HK40043934B - Search method and apparatus, and computer device - Google Patents

Description

Search methods, apparatus and computer equipment

Technical Field

This application relates to the fields of artificial intelligence technology and search query technology, and in particular to a search method, apparatus and computer device.

Background Technology

With the rapid development of science and technology, the internet has become the mainstream platform for information retrieval in today's society. The internet platform contains massive amounts of data, and the emergence of search engines makes it easier for users to find content that meets their needs more quickly within this vast amount of data.

Traditional search engine platforms suggest related terms based on keywords in the search query to help users quickly input their desired words. However, this traditional method is limited to providing auxiliary suggestions based solely on the user's input. As a result, users cannot learn more about the search-related information during the input process, leading to a relatively low amount of information conveyed during the input.

Summary of the Invention

Therefore, it is necessary to provide a search method, apparatus, computer device, and storage medium that can improve the amount of information expressed during the input process, in order to address the above-mentioned technical problems.

A search method, the method comprising:

The query page is displayed; the query page includes an input area.

When a search term is detected in the input area, aggregated suggestions corresponding to the entered search term are displayed in real time on the query page.

The aggregated suggestion content includes multiple index categories and category content related to the search term under each index category; the multiple index categories are index categories supported by the search engine and are related to the query intent of the search term.

A search device, characterized in that the device comprises:

The display module is used to display the query page; the query page displays an input area.

A listening module is used to listen for search terms entered in the input area;

The display module is also used to display aggregated prompts corresponding to the entered search term in real time on the query page when the monitoring module detects that a search term has been entered in the input area.

The aggregated suggestion content includes multiple index categories and category content related to the search term under each index category; the multiple index categories are index categories supported by the search engine and are related to the query intent of the search term.

In one embodiment, the input area is a search box. The display module is also configured to, when a search term is detected being entered into the search box, display a drop-down list below the search box, and in the drop-down list, display the category content related to the search term under the multiple index categories in sequence.

In one embodiment, the display module is further configured to display the category content in the drop-down box in descending order of matching degree between the category content and the user profile, and to display the corresponding index category corresponding to the category content; the user profile refers to the user profile of the user who inputs the search term.

In one embodiment, the display module is further configured to, when the listening module detects a triggering operation on the category content displayed in the aggregated prompt content, redirect to the page corresponding to the triggered category content; wherein the page redirected to corresponds to the index category to which the triggered category content belongs.

In one embodiment, the aggregated prompt content further includes intent-related prompt words associated with the search term. The display module is also configured to, when the monitoring module detects a triggering operation for the intent-related prompt word, redirect to a first search results page; wherein, the first search results page displays the search results obtained using the triggered intent-related prompt word as the search condition.

In one embodiment, the aggregated prompt content further includes a search result viewing entry. The display module is also configured to redirect to a second search result page when the listening module detects a trigger operation on the search result viewing entry; wherein, the second search result page displays the search results obtained using the search term as the search criteria.

A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program performing the following steps:

The query page is displayed; the query page includes an input area.

When a search term is detected in the input area, aggregated suggestions corresponding to the entered search term are displayed in real time on the query page.

The aggregated suggestion content includes multiple index categories and category content related to the search term under each index category; the multiple index categories are index categories supported by the search engine and are related to the query intent of the search term.

A computer-readable storage medium having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

The query page is displayed; the query page includes an input area.

When a search term is detected in the input area, aggregated suggestions corresponding to the entered search term are displayed in real time on the query page.

The aggregated suggestion content includes multiple index categories and category content related to the search term under each index category; the multiple index categories are index categories supported by the search engine and are related to the query intent of the search term.

The aforementioned search method, apparatus, computer equipment, and storage medium, upon detecting a search term entered in the input area of the query page, display in real time aggregated suggestions including index categories supported by the search engine that are relevant to the search intent of the search term, as well as category content related to the search term under each index category. That is, during the search input process, it can suggest both content related to the search term and the index categories supported by the search engine, thereby enabling the user to understand relevant search engine information. Therefore, compared to simply providing auxiliary suggestions for the entered words, it can provide more search-related information and increase the amount of information conveyed.

A search method, the method comprising:

Receive the search terms entered on the query page;

Filter multiple index categories that are relevant to the query intent of the search terms from the index categories supported by the search engine;

Obtain the category content related to the search term under the filtered multiple index categories;

Based on the selected index categories and category content, aggregated suggestion content corresponding to the search term is generated and returned; the aggregated suggestion content is used to indicate that the aggregated suggestion content is displayed on the query page.

A search device, the device comprising:

The receiving module is used to receive the search terms entered on the query page.

The index category filtering module is used to filter multiple index categories that are related to the query intent of the search terms from the index categories supported by the search engine.

The acquisition module is used to acquire category content that is related to the search term under multiple filtered index categories;

The aggregation suggestion module is used to generate and return aggregation suggestion content corresponding to the search term based on the filtered index category and the category content; the aggregation suggestion content is used to indicate that the aggregation suggestion content is displayed on the query page.

In one embodiment, the acquisition module is further configured to recall candidate category content related to the search term under the filtered plurality of index categories; when there are multiple candidate category contents under the same index category, the preset number of candidate category contents are selected from the multiple candidate category contents under the same index category in descending order of text relevance between each candidate category content and the search term, and used as category content for generating aggregated prompt content.

In one embodiment, the acquisition module is further configured to determine the recall method corresponding to each of the selected index categories based on the search term; and for each selected index category, recall the category content related to the search term under the index category according to the recall method corresponding to the index category.

In one embodiment, the aggregation suggestion module is further configured to sort the index categories in descending order of category intent intensity corresponding to the filtered index categories, or sort the category content in descending order of relevance between the category content and the search term; and generate and return aggregate suggestion content corresponding to the search term based on the sorted index categories and the category content under the index categories.

In one embodiment, the aggregation suggestion module is further configured to obtain a user profile of the user who inputs the search term; determine the matching degree between the category content and the user profile; sort the category content in descending order of the matching degree; and generate and return aggregate suggestion content corresponding to the search term based on the sorted index categories and the category content under the index categories.

A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program performing the following steps:

Receive the search terms entered on the query page;

Filter multiple index categories that are relevant to the query intent of the search terms from the index categories supported by the search engine;

Retrieve category content related to the search term from multiple filtered index categories;

Based on the selected index categories and category content, aggregated suggestion content corresponding to the search term is generated and returned; the aggregated suggestion content is used to indicate that the aggregated suggestion content is displayed on the query page.

A computer-readable storage medium having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

Receive the search terms entered on the query page;

Filter multiple index categories that are relevant to the query intent of the search terms from the index categories supported by the search engine;

Retrieve category content related to the search term from multiple filtered index categories;

Based on the selected index categories and category content, aggregated suggestion content corresponding to the search term is generated and returned; the aggregated suggestion content is used to indicate that the aggregated suggestion content is displayed on the query page.

The aforementioned search method, apparatus, computer equipment, and storage medium, upon receiving the input search term, can in real time filter index categories related to the query intent of the search term from the index categories supported by the search engine, and obtain the category content related to the search term under each index category, generate aggregated prompt content, and return it to the terminal. This allows the terminal to provide prompts related to the search term and the index categories supported by the search engine during the input process, thereby enabling the user to know relevant information about the search engine. Therefore, compared to simply providing auxiliary prompts for the input words, it can provide more search-related information and increase the amount of information conveyed.

Attached Figure Description

Figure 1 is an application environment diagram of the search method in one embodiment;

Figure 2 is a flowchart illustrating the search method in one embodiment;

Figure 3 is a schematic diagram of the interface displaying the aggregated prompt content in one embodiment;

Figure 4 is a schematic diagram of the composition of the entity library and the thesaurus in one embodiment;

Figure 5 is a schematic diagram of index category filtering in one embodiment;

Figure 6 is a schematic diagram of a recall method in one embodiment;

Figure 7 is a flowchart illustrating the search method in another embodiment;

Figure 8 is a schematic diagram of the principle of the search method in one embodiment;

Figure 9 is a structural block diagram of the search device in one embodiment;

Figure 10 is a structural block diagram of the search device in another embodiment;

Figure 11 is an internal structure diagram of a computer device in one embodiment;

Figure 12 is an internal structural diagram of a computer device in another embodiment.

Detailed Implementation

To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

The search method provided in this application can be applied to the application environment shown in Figure 1. In this application, terminal 102 communicates with server 104 via a network. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. Server 104 can be a standalone server or a server cluster consisting of multiple servers.

Terminal 102 can display a query page, which includes an input area. When a search term is detected in the input area, terminal 102 can send the detected search term to server 104. Server 104 can then filter multiple index categories related to the search intent of the search term from the index categories supported by the search engine, and retrieve category content related to the search term under the filtered multiple index categories. Server 104 can generate aggregated suggestion content corresponding to the search term based on the filtered index categories and the category content, and return it to terminal 102. Terminal 102 can then display the aggregated suggestion content corresponding to the input search term in real time on the query page.

It should be noted that in the search methods of various embodiments of this application, the search terms and aggregated prompts are displayed in real time, that is, as the input progress of the search terms changes, the displayed aggregated prompts will also change accordingly.

It is understood that the search methods in the various embodiments of this application are equivalent to using artificial intelligence technology to automatically suggest aggregated prompts corresponding to the search terms.

Artificial Intelligence (AI) is the theory, methods, technology, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can react in a way similar to human intelligence. AI studies the design principles and implementation methods of various intelligent machines, enabling them to possess the functions of perception, reasoning, and decision-making.

Artificial intelligence (AI) is a comprehensive discipline encompassing a wide range of fields, including both hardware and software technologies. Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating/interactive systems, and mechatronics. AI software technologies primarily include computer vision, speech processing, natural language processing, and machine learning/deep learning.

It is understood that the search methods in the various embodiments of this application are equivalent to using natural language processing technology in artificial intelligence to identify the query intent of search terms, so as to display multiple index categories and corresponding recalled category content related to the query intent, thereby automatically prompting aggregated prompt content corresponding to the search terms.

Natural Language Processing (NLP) is an important field within computer science and artificial intelligence. It studies the theories and methods for enabling effective communication between humans and computers using natural language. NLP is a science that integrates linguistics, computer science, and mathematics. Therefore, research in this field involves natural language—the language people use in daily life—and thus it has a close relationship with linguistic research. NLP techniques typically include text processing, semantic understanding, machine translation, question answering, and knowledge graphs.

In one embodiment, as shown in Figure 2, a search method is provided. Taking the application of this method to the terminal in Figure 1 as an example, the method includes the following steps:

Step 202: Display the query page; the query page displays an input area.

The query page is where users can enter search terms to trigger a search query.

The input area is the area used to receive input statements.

In one embodiment, the input area may include a search box. In other embodiments, the input area may also include selection buttons, allowing users to enter search terms simply by clicking to select them. In this case, a dedicated search box is not required for user input, and search terms can still be entered.

Step 204: When a search term is detected in the input area, aggregated suggestions corresponding to the entered search term are displayed in real time on the query page. The aggregated suggestions include multiple index categories and category content related to the search term under each index category. The multiple index categories are those supported by the search engine and are related to the query intent of the search term.

It is understandable that aggregated prompt content refers to content composed of multiple prompt messages.

In this embodiment, the aggregated suggestion content includes multiple index categories and category content related to the search term under each index category. It can be understood that the category content under multiple index categories also constitutes multiple different suggestion contents; therefore, it can be called aggregated suggestion content.

It's understandable that for time-sensitive search terms, the aggregated suggestions might differ at different times. For example, if the search term is a news-sensitive keyword, during that time period, the aggregated suggestions should display the most relevant news result; after the time period expires, other relevant results can be displayed.

The multiple index categories displayed in the aggregated suggestion content are index categories supported by the search engine and are related to the query intent of the search terms.

The index category refers to the category or type that can be indexed for queries. The index categories supported by a search engine are those categories that provide the search results when a query is performed within the search engine. In other words, content related to the search terms can be found from these index categories.

In one embodiment, the index categories include at least one of the following: sub-application (i.e., mini-program) index category, public account index category, tool index category, multimedia content index category (e.g., audio, video, and articles), information index category (e.g., news), entertainment index category (e.g., music, games, and novels), and question and answer index category.

In this context, a child application is an application that can run within the environment provided by the parent application. The parent application is the application that hosts the child application, providing the environment for its implementation. The parent application is a native application, which can run directly on the operating system. Child applications can run based on the parent application without being downloaded, and are considered lightweight applications. For example, WeChat is the parent application, and WeChat Mini Programs are child applications running within the native parent application of WeChat.

In one embodiment, a sub-application may include sub-applications in various scenarios such as instant messaging sub-applications, mobile payment sub-applications, and game sub-applications. As long as it meets the characteristic of being a lightweight application that runs on the parent application without the need for downloading or installation, it can be called a sub-application.

A public account refers to an application account on a public platform, enabling communication with specific groups within the platform through text, images, voice, and video. In one embodiment, a public account can be a public service account on a social media platform (such as a public service account on WeChat), or a service account on some third-party payment platforms, etc.

It should be noted that the indexable categories supported by a search engine can be determined according to the search engine's own needs, and different search engines support different indexable categories.

The following examples illustrate the index categories supported by search engines. Suppose a search engine supports sub-application index categories. When searching for term A, if the sub-application index category is relevant to the query intent of term A, then content related to term A can be found within that sub-application index category; that is, sub-applications related to term A can be retrieved. Similarly, if a search engine supports public account index categories, and this category is relevant to the query intent of term A, then content related to term A can be found within that category; that is, public accounts related to term A can be retrieved.

Specifically, users can enter search terms in the input area. The terminal can monitor input events in the input area in real time. When the input event of a user entering a search term is detected, the terminal sends the detected search term to the server. The server can then filter multiple index categories related to the search intent from the index categories supported by the search engine, and retrieve category content related to the search term under the filtered index categories. Based on the filtered index categories and the corresponding retrieved category content, the server generates aggregated suggestion content corresponding to the search term and returns it to the terminal. The terminal can then display the aggregated suggestion content corresponding to the entered search term in real time on the query page.

In one embodiment, the server can pre-build different types of databases (e.g., entity databases, news terminologies, question-and-answer terminologies, and pattern terminologies) based on aggregate analysis. Then, based on the built databases, it filters multiple index categories related to the query intent of the search term from the index categories supported by the search engine. Furthermore, based on the built databases, it retrieves category content related to the search term from the filtered multiple index categories. In other embodiments, the server can also directly perform intent recognition analysis to determine multiple index categories related to the query intent of the search term, without needing to pre-build different types of databases based on aggregate analysis. The server can also mine the mapping relationship between search terms and category content based on offline big data to retrieve category content, or directly use the search engine to retrieve category content. This is not limited to any particular approach.

In one embodiment, the server can generate a content aggregation card and return it to the terminal, displaying aggregated suggestions within the content aggregation card. The terminal can then display the content aggregation card at a preset location on the search page to show the aggregated suggestions corresponding to the entered search terms.

It should be noted that the aggregated suggestion content includes, but is not limited to, multiple index categories and category content related to the search terms under each index category. In other words, the aggregated suggestion content can also include other content, thus aggregating more information.

In one embodiment, the server can sort the recalled category content, generate aggregated prompt content based on the sorted category content and the corresponding index category, and return it to the terminal.

The aforementioned search method, upon detecting a search term entered in the input area of the query page, displays in real-time aggregated suggestions, including index categories supported by the search engine that are relevant to the search intent of the term, as well as category content related to the search term within each index category. In other words, during the search input process, it provides suggestions not only for content related to the search term but also for the index categories supported by the search engine, thereby informing the user of relevant search engine information. Therefore, compared to simply providing auxiliary suggestions for the entered words, it offers more search-related information and increases the amount of information conveyed.

In one embodiment, the input area is a search box. In step 204, when a search term is detected being entered in the input area, aggregated suggestions corresponding to the entered search term are displayed in real time on the query page. This includes: when a search term is detected being entered in the search box, a drop-down list is displayed below the search box, and the category content related to the search term under the multiple index categories is displayed sequentially in the drop-down list.

In this context, "displaying in sequence" refers to arranging the category content related to the search term within multiple index categories in a specific order. It's important to note that "displaying in sequence" here refers to the order in which the category content is arranged across different index categories.

In one embodiment, the category content displayed under different index categories can be sorted according to at least one of the following: the matching degree between the category content and the user profile, the category intent strength corresponding to the index category to which the category content belongs, and the relevance between the category content and the search term.

It is understandable that the number of category contents displayed under the same index category can be a preset number. The preset number can be at least one. The category contents displayed under the same index category can be a preset number of category contents selected according to the order of textual relevance between the category contents and the search term from high to low. The preset number can be understood as follows: when the preset number is 1, the category contents displayed under the same index category have the highest textual relevance to the search term.

It should be noted that the category content is displayed according to the corresponding index category, that is, the category content of the same index category is displayed according to the corresponding index category.

In the above embodiments, the aggregated prompts are generated after sorting the category content, which improves the accuracy and reference value of the prompts.

In one embodiment, displaying the category content related to the search term under the multiple index categories in the drop-down box in sequence includes: displaying the category content in the drop-down box in descending order of matching degree between the category content and the user profile, and displaying the corresponding index category corresponding to the category content.

The user profile refers to the user profile of the user who inputs the search term. Multiple index categories are those index categories supported by the search engine that are related to the query intent of the search term.

In one embodiment, the user who enters the search term can be a user with a logged-in account or a user selected to perform the search operation.

Specifically, the server can determine the matching degree between category content related to the search term and the user profile under multiple index categories. Then, it sorts the category content in descending order of matching degree and returns the sorted category content and its corresponding index category to the terminal. The terminal can then display the category content in a drop-down menu, ordered from highest to lowest matching degree between the category content and the user profile, and display the corresponding index category for each category content.

It is understood that the order in which the index categories are arranged is consistent with the order in which the content of the categories under the index categories is arranged.

In one embodiment, the server can perform semantic analysis on category content using a pre-trained matching model, extract semantic features, and then calculate the matching degree between the extracted semantic features and the user profile to obtain the matching degree between each category of content and the user profile.

This is understandable, as it allows for different users to receive different priority levels of category content for the same search term, thus enabling personalized suggestions.

To better illustrate personalized suggestions, let's look at an example. For instance, for the same search term "**live stream," the results would include public accounts, mini-programs, tech articles describing the technology used to enable the live stream, and corresponding live stream video footage. If the user is a tech professional (e.g., someone who frequently browses tech blogs or reviews live stream-related technical materials), the tech article describing the technology used to enable the live stream would appear first in the dropdown menu, as it best matches their profile. If the user is a professional live streamer, the mini-program would appear before the tech article. Conversely, if the user is a casual live streamer, the mini-program or live stream video footage would appear first, with the tech article last.

In the above embodiments, the category content is sorted by the matching degree between the category content under different index categories and the user profile, so that personalized prompts can be given to different users for the same search term during the input process. That is, the priority of the category content prompted to different users is different, which improves the accuracy of prompts during the input process.

In one embodiment, displaying the category content related to the search term under the plurality of index categories in the drop-down box in sequence includes: displaying the plurality of index categories and the category content under the plurality of index categories in the drop-down box according to the category intent intensity corresponding to the plurality of index categories from high to low.

Category intent strength is used to characterize the degree of correlation between the index category and the query intent of the search term.

In one embodiment, the category intent strength can be determined based on the category intent score. That is, the category intent strength can be the category intent score itself, or it can be calculated based on the category intent score.

Specifically, the server can analyze the relevance between multiple index categories and the query intent of search terms to obtain the category intent strength corresponding to each index category. Then, it sorts the category content under each index category in descending order of category intent strength and returns the sorted category content and its corresponding index category to the terminal. The terminal can then display multiple index categories and their category content in a drop-down menu, ordered from highest to lowest category intent strength.

This is understandable, as it allows the system to recommend the most relevant index categories and corresponding content to users, thereby improving recommendation accuracy.

In one embodiment, displaying the category content related to the search term under the plurality of index categories in the drop-down list in sequence includes: displaying the category content in the drop-down list according to the order of high to low relevance between the category content and the search term, and displaying the corresponding index category corresponding to the category content. The relevance is determined based on at least one of the following: the category intent strength of the index category corresponding to the category content, the textual relevance between the category content and the search term, and the timeliness of the category content.

Among them, relevance is used to characterize the correlation between category content and search terms.

It should be noted that relevance can be used to characterize the textual relevance between category content and search terms, or it can be used to characterize the overall relevance between category content and search terms.

In one embodiment, relevance can be determined based on at least one of the following: the category intent strength of the index category corresponding to the category content, the textual relevance between the category content and the search term, and the timeliness of the category content.

In one embodiment, the server can score the text relevance between category content and the search term to obtain a text relevance score, score the timeliness of category content to obtain a content timeliness score, and obtain the category intent score of the index category corresponding to the category content. Then, the server can perform a weighted summation of the text relevance score, content timeliness score, and category intent score to obtain a comprehensive score as the relevance between the category content and the search term.

It's understandable that when scoring based on timeliness, content from categories with higher timeliness scores higher than content from categories with lower timeliness. For example, in the case of news content, the most recent news matching the search term scores higher than outdated news.

Specifically, the server can sort the content of each category in descending order of relevance, and return the sorted category content and its corresponding index category to the terminal. The terminal can then display the category content in a drop-down menu, in descending order of relevance between the category content and the search term, and display the corresponding index category for each category content.

In one embodiment, the aggregated prompt content further includes at least one of intent-related prompts and a search result viewing entry point. It can be understood that when generating a content aggregation card, in addition to displaying multiple index categories and corresponding recalled category content, the content aggregation card may also display at least one of intent-related prompts and a search result viewing entry point.

The intent-related prompts are prompts related to the query intent.

In one embodiment, after the terminal sends the search term to the server, in addition to filtering the index categories related to the query intent of the search term from the index categories supported by the search engine and performing the corresponding category content recall process, the server can also search for intent-related prompt words related to the query intent.

In one embodiment, the intent-related prompt may include at least one of related keywords and extended keywords. Related keywords are words that include keywords from the search term. That is, a portion of the related keywords is a keyword from the search term. Extended keywords are keywords that are not included in the search term but are relevant to the query intent. That is, extended keywords do not include keywords from the search term.

It is understandable that both native keywords and extended keywords can serve as intent-related suggestions when users enter search terms, thus assisting their search queries.

In one embodiment, the server can also generate a search result viewing entry point. This entry point directs users to the search results page to view the complete search results.

In one embodiment, the input area is a search box. When a search term is detected in the input area, a content aggregation card can be displayed at a preset position below the search box to show aggregated suggestions corresponding to the entered search term in real time.

Figure 3 is a schematic diagram of the interface displaying aggregated prompts in one embodiment. Referring to Figure 3, the dashed box 302 shows the search box. Enter the search term "WeChat Open Class" in the search box. It can be understood that the content in dashed box 304 in Figure 3 is the aggregated prompt content displayed in the content aggregation card. The "Official Accounts," "Mini Programs," "Articles," and "Videos" in dashed box 306 are index categories supported by the search engine that are related to the query intent of the search term. The content following each category is the category content related to the search term under that index category. For example, "WeChat Open Class" after the "Official Accounts" index category is the retrieved public account information related to the search term "WeChat Open Class." That is, a public account called "WeChat Open Class" is found through the search. Similarly, as shown in Figure 3, a search reveals the existence of the "WeChat Open Class+" mini program, the article "***WeChat Open Class PRO Speech," and the video "2020 WeChat Open Class." In Figure 3, "Related Searches" represents search results for keywords that match the query intent. "WeChat Open Class PRO" and "WeChat Short Content" following "Related Searches" are intent-related keywords. "WeChat Open Class PRO" is a related keyword that includes the search term "WeChat Open Class." "WeChat Short Content" is an expanded keyword that does not include the search term "WeChat Open Class," but is related to the query intent. "Explore More Search Results" in Figure 3 is the entry point to view search results. This means the search results can be viewed at the bottom of the content aggregation card.

In the above embodiments, in addition to multiple strongly related categories and category content, the aggregated suggestion content also displays intent-related suggestion words and/or search result viewing entry points. Therefore, intelligent suggestions using intent-related suggestion words can shorten the user's input path and improve search efficiency. Furthermore, when displaying the search result viewing entry point, if the content suggested in the aggregated suggestion content (category content and intent-related suggestion words) does not meet the user's needs, the user can enter the search results page to see the complete search results, thus improving the search success rate.

In one embodiment, the aggregated suggestion content further includes intent-related suggestion words associated with the search term. The method also includes: when a triggering operation on category content displayed in the aggregated suggestion content is detected, redirecting to the page corresponding to the triggered category content; wherein the redirected page matches the index category to which the triggered category content belongs.

It's understandable that users can trigger the displayed aggregated prompts (such as by clicking or pressing them).

Specifically, users can trigger operations on the category content displayed in the aggregated prompts. When the terminal detects a trigger operation on the category content displayed in the aggregated prompts, it will redirect to the page corresponding to the triggered category content.

In one embodiment, when a terminal detects a triggering action on category content displayed in an aggregated prompt, it can request the page information to be navigated to from the server. The server can then determine the page information corresponding to the triggered category content and send the page information back to the terminal. The terminal then navigates from the query page to the page corresponding to the triggered category content based on this page information.

It's understandable that the page redirected to matches the index category to which the triggered content belongs. For example, if the triggered content is from a sub-application index category, then the page redirected to will match that sub-application index category. Referring to Figure 3, assuming the triggered content is from the "WeChat Open Class+" category under the Mini Program index category, then the user will be redirected to the "WeChat Open Class+" Mini Program page.

In one embodiment, the page corresponding to the triggered category content can include either a landing page for the triggered category content or an account entry page. A landing page, also known as a landing page or onboarding page, is a webpage displayed to users after a search engine result is obtained on the internet. An account entry page is the page used to log in to an account.

In the above embodiments, during the input process, aggregated prompts suggest category content related to the query intent. After the user triggers the suggested category content, they are directly redirected to the page corresponding to the triggered category content, enabling them to enter the specific page more quickly, without having to enter the search results page first and then select the specific page, thus improving search efficiency.

In one embodiment, the aggregated prompt content further includes a search result viewing entry. The method also includes: when a triggering operation targeting an intent-related prompt word is detected, redirecting to a first search result page; wherein the first search result page displays search results retrieved using the triggered intent-related prompt word as the search condition.

Specifically, users can also trigger (e.g., click) intent-related prompts within the aggregated prompt content. When the terminal detects a trigger action targeting an intent-related prompt, it can send a search request to the server using that prompt as a search criterion. The server can then perform a search using the triggered intent-related prompt as the search criterion and generate a first search results page, which is then sent back to the terminal. The terminal can then navigate from the query page to the first search results page.

In the above embodiments, through intelligent suggestions based on intent association keywords, users can directly select intent association keywords to perform search queries without needing to input complete sentences, thus shortening the user's input path and improving search efficiency.

In one embodiment, the method further includes: when a trigger operation targeting the search results viewing entry is detected, redirecting to a second search results page; wherein, the second search results page displays the search results obtained using the search term as the search criteria.

Specifically, users can also trigger (for example, by clicking) the search results viewing entry in the aggregated prompt content. When the terminal detects a trigger action targeting the search results viewing entry, it can send a search request to the server using the entered search term as the search criteria. The server can then perform a search using the search term as the search criteria and generate a second search results page, which is then sent back to the terminal. The terminal can then jump from the query page to the second search results page.

In the above embodiments, when the suggested content does not meet the requirements, the search results viewing entry can be directly triggered to view the complete search results, thereby improving the search success rate.

In one embodiment, the multiple index categories in the aggregated prompt content are obtained by performing category intent analysis on the search terms to obtain the category intent score corresponding to each index category supported by the search engine, and then filtering the index categories from the index categories supported by the search engine based on the category intent score corresponding to each index category.

Among them, the category intent score is used to characterize the degree of relevance between the query intent of the search term and the index category.

It's understandable that the relevance between the search term's query intent and the indexed category is positively correlated with the category intent score. A higher category intent score indicates a higher relevance between the search term's query intent and that indexed category; conversely, a lower category intent score indicates a lower relevance.

Specifically, after the terminal sends the search terms to the server, the search terms instruct the server to perform category intent analysis on the search terms, obtain the category intent score corresponding to each index category supported by the search engine, and then select the index categories from the index categories supported by the search engine based on the category intent score. That is, based on the category intent score, the server selects index categories related to the query intent of the search terms from the index categories supported by the search engine.

In one embodiment, after the terminal sends the search term to the server, the search term is used to instruct the server to input the search term into a pre-trained classification model to predict the score corresponding to the search term under each index category supported by the search engine, which is used as the category intent score corresponding to each index category supported by the search engine.

In one embodiment, different types of databases are pre-built in the server. Each database pre-stores category content for at least one index category, with category content under the same index category stored in the same database. Each category of content has a corresponding index category annotation. It can be understood that the index category represented by the index category annotation is the index category supported by the search engine.

In one embodiment, some categories in the database may already have indexed category tags, while others may not. In this case, the appropriate indexed category tags can be added to the categories that do not have them by manual addition, analysis of user behavior logs, and pattern string matching.

In one embodiment, the server can extract the index categories clicked by the user during the historical search process from the user behavior log, and use the clicked index categories as the index category tags for the category content searched by the user.

In one embodiment, pattern string matching processing refers to extracting a fixed query pattern string from the category content that has not been indexed by a target tag, then matching the fixed query pattern string with the category content that has been indexed by a target tag, and using the target tag corresponding to the matched category content as the target tag of the category content that has not been indexed by a target tag.

The server can pre-train a classification model using machine learning based on preset category content and corresponding index category annotations stored in the database. Then, the server can input the search terms sent by the terminal into this pre-trained classification model to predict the score of the search term under each index category supported by the search engine, thus obtaining the category intent score corresponding to each index category supported by the search engine.

In one embodiment, the classification model can be a classification model based on FastText or XGBoost.

In one embodiment, pre-built databases of different categories may include entity libraries (i.e., classification databases) and vocabularies.

In this context, an entity is an objectively existing and distinguishable thing. In the context of databases, an entity often refers to a collection of things of a certain kind. Each individual data object of a category is called an entity. An entity database can be understood as a pre-built categorized database constructed by aggregating and classifying content according to preset categories. Category content belonging to the same index category is aggregated in the same entity database.

It is understandable that the data content in the entity database is the name.

In one embodiment, the entity library may include an account entity library and a name entity library. For example, the names of public accounts and mini-program accounts can each form two separate account entity libraries, and music titles and novel titles can also form corresponding name entity libraries. It is understood that the aggregation and categorization of the entity libraries can be set according to actual needs and is not limited thereto. It is also understood that each type of entity library can consist of one or more entities.

Taking Figure 4 as an example, in the example in Figure 4, "Account Library", "Music Library" and "Game Name" are the names of different entity libraries (i.e., different index categories), while "WeChat Sports" and "Honor of Kings" are the specific names stored in the entity libraries (i.e., category content).

A thesaurus is a pre-defined collection of content from pre-defined categories. Content belonging to the same index category is grouped into the same thesaurus.

In one embodiment, the thesaurus may include at least one of a pattern thesaurus, an information thesaurus, a question-and-answer thesaurus, a tool thesaurus, and an entry thesaurus. It is understood that each type of thesaurus may be one or more.

In one embodiment, the server can extract fixed query pattern strings based on manually pre-configured fixed query pattern strings. Alternatively, it can perform aggregation analysis on preset category content to extract fixed query pattern strings, thereby obtaining a pattern vocabulary based on these strings. It can be understood that fixed query pattern strings support matching methods such as regular expressions (i.e., satisfying general regular expression matching expressions). As shown in Figure 4, both "Weather Pattern" and "Route Query" have fixed query pattern strings, so corresponding vocabulary lists can be built for them. For example, the fixed query pattern string for the Weather Pattern is "**Weather". Therefore, when querying different cities or regions, only the content in ** needs to be replaced. Thus, it has a fixed query pattern string, and a corresponding vocabulary list can be created for the Weather Pattern.

In one embodiment, the server can perform cluster analysis on the titles and content of trending news to obtain news keywords, or obtain manually edited news keywords, and then construct a news thesaurus based on the extracted news keywords. As shown in Figure 4, the news keyword extracted from trending news is "COVID-19".

In one embodiment, the server can extract keywords from the question-and-answer result title and content to obtain question-and-answer keywords, or obtain manually edited question-and-answer keywords, and then construct a question-and-answer thesaurus based on the obtained question-and-answer keywords. As shown in Figure 4, extracting question-and-answer keywords from medical questions yields "What to do about a cold".

In one embodiment, the server can perform aggregation analysis on entries in the thesaurus to obtain an entry thesaurus. As shown in Figure 4, "Scenic Spots Encyclopedia" and "Biographies Encyclopedia" are the results of the aggregation analysis, where "Forbidden City" is the specific entry content.

In the above embodiments, by performing category intent analysis on search terms and filtering index categories from those supported by the search engine based on the category intent scores corresponding to each index category, index categories that match the user's query intent can be filtered more accurately, improving the accuracy and relevance of suggestions during the input process.

In one embodiment, the multiple index categories in the aggregated prompt content are the first preset number of index categories whose category intent scores are greater than or equal to a preset intent score threshold after sorting the category intent scores corresponding to each index category supported by the search engine.

It is understood that in this embodiment, the search term is also used to instruct the server to sort the category intent scores corresponding to each index category supported by the search engine, and to filter out the top preset number of index categories whose category intent scores are greater than or equal to a preset intent score threshold. The preset number can be set according to actual needs.

Specifically, the server performs category intent analysis on the search terms, obtaining the category intent score corresponding to each index category supported by the search engine. Then, it sorts these category intent scores from highest to lowest, and selects the top preset number of index categories whose category intent scores are greater than or equal to a preset threshold. It should be noted that the selected index categories must simultaneously meet two conditions: their category intent score must be greater than or equal to the preset threshold, and they must be ranked within the top preset positions.

Figure 5 is a schematic diagram of index category filtering in one embodiment. Referring to Figure 5, for the search term "WeChat Open Class", the categories are sorted in descending order of their category intent scores under the index categories supported by the search engine. Assuming that the preset intent score threshold is 0.5, and the first 4 index categories (i.e., the preset quantity is 4), then the 4 index categories "Public Accounts", "Mini Programs", "News", and "Videos" can be filtered out.

In the above embodiments, after sorting the category intent scores corresponding to each index category supported by the search engine, the top preset number of index categories with category intent scores greater than or equal to a preset intent score threshold are selected. This enables more accurate and faster selection of index categories that match the user's query intent, improving the accuracy and relevance of prompts during the input process.

In one embodiment, the category content is the prompt content related to the search term under each of the selected index categories, which is retrieved according to the retrieval method corresponding to each of the selected index categories.

It is understood that in this embodiment, the search term is also used to instruct the server to determine the recall method corresponding to each of the selected index categories based on the search term, and to recall the category content related to the search term under each of the selected index categories according to the corresponding recall method.

It should be noted that the server can return all the recalled category content, or it can perform further filtering on the recalled category content to select category content with higher relevance to the search terms, and then return the filtered category content to the terminal.

Specifically, after filtering the index categories, the server can also determine the corresponding recall method for each filtered index category based on the search terms. In other words, different index categories can use different recall methods; that is, the corresponding category content is retrieved using the recall method specific to each category.

In one embodiment, the recall method includes at least one of entity database recall, vocabulary recall, and search engine recall.

Among them, the entity database recall method is used to indicate the recall of category content related to the search term from the entity database corresponding to the index category.

The term recall method is used to indicate the retrieval of category content related to the search term from the terminology corresponding to the indexed category.

Search engine recall methods are used to instruct the retrieval of category content related to the search term from the search engine.

It's understandable that a corresponding recall method can be pre-set for each index category. An index category can correspond to at least one recall method. When an index category corresponds to multiple recall methods, one method can be randomly selected for recall processing, or recall methods can be selected in a preset order. If the previous recall method fails, the next recall method will be used to retrieve the relevant prompts for the search term under that index category.

It should be noted that a single recall method can include multiple sub-recall methods. There is also a priority order among the sub-recall methods under the same recall method.

In one embodiment, the entity database retrieval method may include two sub-retrieval methods: direct retrieval and prefix tree creation retrieval. Direct retrieval refers to directly retrieving data from the entity database. Prefix tree creation retrieval refers to creating a prefix tree based on the names in the entity database to which the indexed category belongs, and then matching the search terms with the prefixes of the names in the prefix tree for retrieval.

In one embodiment, the vocabulary-based recall method may include a sub-recall method that retrieves responses by matching them with a fixed query pattern string in the pattern vocabulary, and a sub-recall method that retrieves responses from a specific vocabulary. It can be understood that recalling responses by matching them with a fixed query pattern string in the pattern vocabulary is equivalent to fuzzy matching recall, which can improve recall efficiency, and therefore is preferred over recalling responses from a specific vocabulary.

In other embodiments, recall processing can also be carried out through other recall methods, such as offline aggregation result recall.

Figure 6 illustrates a recall method in one embodiment. Referring to Figure 6, each index category can correspond to one or more recall methods. For example, index categories such as public accounts/mini-programs/music/games correspond to entity database recall and search engine recall. The entity database recall method further includes two sub-recall methods: direct recall and trie tree recall. As shown in Figure 6, the direct recall method is preferred, directly recalling the corresponding results (i.e., category content) from the entity database. If direct recall fails, the trie tree recall method is used. If the corresponding results still cannot be recalled, an inverted index can be built for the entity database name, and the search engine system can be used for recall, i.e., the search engine recall method is used.

For example, when querying index categories for tools, the corresponding recall methods include terminology recall and search engine recall. The terminology recall method includes sub-recall methods that match results with fixed query pattern strings in the pattern terminology, and sub-recall methods that retrieve results from specific tool terms. As shown in Figure 6, the sub-recall method that matches results with fixed query pattern strings in the pattern terminology is used first. If the corresponding results cannot be retrieved directly, the corresponding results are retrieved from the tool terminology. If the corresponding results still cannot be retrieved, an inverted index is created for the tool category name, and the results are retrieved through the search engine system; that is, the search engine recall method is used.

Inverted indexes use documents as the key index. The record entries corresponding to the keywords in the table record all documents containing that word or phrase. Each entry is a word segment, recording the document's identifier and the position of the character within that document. It can be understood that creating an inverted index for an entity database name is equivalent to using the entity database name as the key index. Similarly, creating an inverted index for a tool category name is equivalent to using the tool category name as the key index. The specific index categories in Figure 6 will not be explained in detail here, as the retrieval principle is the same as that of the already explained index categories.

In the above embodiments, different index categories can adopt different recall methods, which improves recall accuracy and recall efficiency.

As shown in Figure 7, in one embodiment, a search method is provided. Taking the application of this search method to the server in Figure 1 as an example, the method includes the following steps:

Step 702: Receive the search terms entered on the query page.

Step 704: Select multiple index categories from the index categories supported by the search engine that are related to the query intent of the search term.

Step 706: Obtain category content related to the search term from the multiple filtered index categories.

Step 708: Based on the selected index categories and category content, generate and return aggregated suggestion content corresponding to the search terms; the aggregated suggestion content is used to indicate whether to display aggregated suggestion content on the query page.

In one embodiment, the aggregated prompt content further includes at least one of intent-related prompts associated with the search term and a search result viewing entry; wherein, the intent-related prompts are used to redirect to the search results page obtained by searching with the intent-related prompts as search criteria after being triggered; and the search results viewing entry is used to redirect to the search results page obtained by searching with the search term as search criteria after being triggered.

In one embodiment, filtering multiple index categories related to the query intent of a search term from the index categories supported by the search engine includes: performing category intent analysis on the search term to obtain the category intent score corresponding to each index category supported by the search engine; filtering the index categories supported by the search engine based on the category intent score corresponding to each index category; and ensuring that the category intent score corresponding to the filtered index categories is higher than the category intent score corresponding to the unfiltered index categories.

In one embodiment, performing category intent analysis on search terms to obtain category intent scores corresponding to each index category supported by the search engine includes: inputting the search term into a pre-trained classification model, predicting the score corresponding to the search term under each index category supported by the search engine, and using this score as the category intent score corresponding to each index category supported by the search engine.

In one embodiment, the server can perform machine learning training in advance based on the preset category content and corresponding index category annotations stored in the database to train a classification model.

In one embodiment, filtering the index categories supported by the search engine based on the category intent score corresponding to each index category includes: sorting the category intent scores corresponding to each index category supported by the search engine, and filtering the top preset number of index categories whose category intent scores are greater than or equal to a preset intent score threshold.

In one embodiment, step 706, obtaining category content related to the search term under the filtered plurality of index categories, includes: recalling candidate category content related to the search term under the filtered plurality of index categories; when there are multiple candidate category contents under the same index category, selecting a predetermined number of candidate category contents from the multiple candidate category contents under the same index category in descending order of text relevance between each candidate category content and the search term, as category content for generating aggregated prompt content.

The preset number can be one or at least two. It should be noted that "multiple" in the embodiments of this application means at least two. Text relevance refers to the relevance between the text of the category content and the text of the search term.

It is understandable that when there are multiple candidate category contents under the same index category, the first preset number of candidate category contents are selected in descending order of text relevance as the category contents used to generate aggregated suggestion content. This can more accurately recommend the category contents that users want to search for, thus improving the accuracy of recommendations.

In one embodiment, generating and returning aggregated suggestion content corresponding to the search term based on the selected index categories and category content includes: sorting the index categories in descending order of category intent intensity; and generating and returning aggregated suggestion content corresponding to the search term based on the sorted index categories and category content under the index categories.

Specifically, the server can analyze the correlation between multiple index categories and the query intent of search terms to obtain the category intent strength corresponding to each index category. Then, it sorts the category content under the index category in descending order of category intent strength. Based on the sorted index categories and the category content under the index categories, it generates aggregated suggestion content corresponding to the search term and returns the aggregated suggestion content to the terminal.

In one embodiment, generating and returning aggregated suggestion content corresponding to the search term based on the selected index categories and category content includes: sorting the category content in descending order of relevance between the category content and the search term; generating and returning aggregated suggestion content corresponding to the search term based on the sorted index categories and the category content under the index categories.

Specifically, the server can analyze the relevance between category content and search terms based on at least one of the following: the category intent strength of the index category corresponding to the category content, the textual relevance between the category content and the search term, and the timeliness of the category content. The server can sort the category content in descending order of relevance between the category content and the search term. Based on the sorted index categories and the category content under each index category, the server can generate aggregated suggestion content corresponding to the search term and return the generated aggregated suggestion content to the terminal.

In one embodiment, the server can score the text relevance between category content and the search term to obtain a text relevance score, score the timeliness of category content to obtain a content timeliness score, and obtain the category intent score of the index category corresponding to the category content. Then, the server can perform a weighted summation of the text relevance score, content timeliness score, and category intent score to obtain a comprehensive score as the relevance between the category content and the search term.

In one embodiment, based on the selected index categories and category content, aggregated suggestion content corresponding to the search term is generated and returned, including: obtaining the user profile of the user who inputs the search term; determining the matching degree between the category content and the user profile; sorting the category content in descending order of matching degree; and generating and returning aggregated suggestion content corresponding to the search term based on the sorted index categories and the category content under the index categories.

The user profile refers to the user profile of the user who inputs the search term. Multiple index categories are those index categories supported by the search engine that are related to the query intent of the search term.

In one embodiment, the user who enters the search term can be a user with a logged-in account or a user selected to perform the search operation.

Specifically, the server can determine the matching degree between category content related to the search term and the user profile under multiple index categories. Then, it sorts the category content in descending order of matching degree and returns the sorted category content and its corresponding index category to the terminal. The terminal can then display the category content in a drop-down menu, ordered from highest to lowest matching degree between the category content and the user profile, and display the corresponding index category for each category content.

It is understood that the order in which the index categories are arranged is consistent with the order in which the content of the categories under the index categories is arranged.

In one embodiment, the server can perform semantic analysis on category content using a pre-trained matching model, extract semantic features, and then calculate the matching degree between the extracted semantic features and the user profile to obtain the matching degree between each category of content and the user profile.

This is understandable, as it allows for different users to receive different priority levels of category content for the same search term, thus enabling personalized suggestions.

In the above embodiments, the category content is sorted by the matching degree between the category content under different index categories and the user profile, so that personalized prompts can be given to different users for the same search term during the input process. That is, the priority of the category content prompted to different users is different, which improves the accuracy of prompts during the input process.

In one embodiment, recalling category content related to the search term under multiple filtered index categories includes: determining the recall method corresponding to each filtered index category based on the search term; and for each filtered index category, recalling the category content related to the search term under the index category using the corresponding recall method.

In one embodiment, the recall method includes at least one of entity database recall, vocabulary recall, and search engine recall.

In one embodiment, the server can sort the recalled category content, generate aggregated prompt content based on the sorted category content and the corresponding index category, and return it to the terminal.

In one embodiment, the server can sort the category content under each index category according to the category intent score, from highest to lowest. The server can also score the category content based on at least one of the textual relevance between the recalled category content and the search terms, and the timeliness of the category content, and sort the category content according to the score, from highest to lowest. It is understood that the server can also combine the score and the category intent score (e.g., weighted summation) to determine a comprehensive score, and sort the category content according to the comprehensive score, from highest to lowest. This is not limited.

It's understandable that when scoring based on timeliness, content in categories with higher timeliness scores higher than content in categories with lower timeliness. For example, in the case of news content, the most recent news items matching the search terms score higher than outdated news items.

The aforementioned search method, apparatus, computer equipment, and storage medium, upon receiving the input search term, can in real time filter index categories related to the query intent of the search term from the index categories supported by the search engine, as well as recall category content related to the search term under each index category, generate aggregated prompt content, and return it to the terminal. This allows the terminal to provide both content related to the search term and the index categories supported by the search engine during the input process, thereby enabling the user to be aware of search engine-related information. Therefore, compared to simply providing auxiliary prompts for the input words, it can provide more search-related information and increase the amount of information conveyed.

Figure 8 is a schematic diagram illustrating the principle of the search method in one embodiment. Referring to Figure 8, the user enters a search term in the search box on the terminal's query page. The terminal sends the search term to the server, which performs category intent analysis on the search term based on pre-established databases of different categories (e.g., entity database, news terminology for news aggregation, pattern terminology, and question-and-answer terminology). Based on the analysis results, the server filters index categories related to the query intent of the search term. The server can determine the recall method corresponding to the filtered index category (different index categories have different recall methods). Based on the determined recall method, and again combining it with the established databases (e.g., entity database, news terminology for news aggregation, pattern terminology, and question-and-answer terminology), the server recalls category content (i.e., recalls index category results) under the selected index category. Then, the server can sort the recalled category content, combine it to obtain aggregated prompt content, and generate a content aggregation card. The terminal can display this content aggregation card in real time below the search box.

It should be understood that although the steps in the flowchart above are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps.

In one embodiment, as shown in FIG9, a search device is provided. This device can be a software module, a hardware module, or a combination of both, integrated into a computer device. Specifically, the device includes a display module 902 and a listening module 904, wherein:

Display module 902 is used to display the query page; the query page displays an input area.

The listening module 904 is used to listen for search terms entered in the input area.

The display module 902 is also used to display aggregated suggestions corresponding to the entered search term in real time on the query page when the listening module 904 detects that a search term has been entered in the input area.

The aggregated suggestion content includes multiple index categories and category content related to the search terms under each index category; the multiple index categories are index categories supported by the search engine that are related to the query intent of the search terms.

In one embodiment, the input area is a search box. The display module 902 is further configured to, when a search term is detected being entered into the search box, display a drop-down list below the search box, and in the drop-down list, display the category content related to the search term under the plurality of index categories in sequence.

In one embodiment, the display module 902 is further configured to display the category content in the drop-down box in descending order of matching degree between the category content and the user profile, and to display the corresponding index category corresponding to the category content; the user profile refers to the user profile of the user who inputs the search term.

In one embodiment, the display module 902 is further configured to, when the listening module 904 detects a trigger operation on the category content displayed in the aggregated prompt content, jump to the page corresponding to the triggered category content; wherein the page to which the jump is made conforms to the index category to which the triggered category content belongs.

In one embodiment, the aggregated prompt content further includes at least one of intent-related prompts associated with the search term and a search result viewing entry; wherein, intent-related prompts are prompts associated with the query intent.

In one embodiment, the aggregated prompt content further includes intent-related prompt words associated with the search term. The display module 902 is also configured to, when the listening module 904 detects a triggering operation for the intent-related prompt word, redirect to the first search results page; wherein, the first search results page displays the search results obtained using the triggered intent-related prompt word as the search condition.

In one embodiment, the aggregated prompt content also includes a search result viewing entry. The display module 902 is further configured to, when the listening module 904 detects a trigger operation on the search result viewing entry, redirect to a second search result page; wherein, the second search result page displays the search results obtained using the search term as the search criteria.

In one embodiment, the multiple index categories in the aggregated prompt content are obtained by performing category intent analysis on the search terms to obtain the category intent score corresponding to each index category supported by the search engine, and then filtering the index categories from the index categories supported by the search engine based on the category intent score corresponding to each index category.

In one embodiment, the category content is the prompt content related to the search term retrieved under each of the selected index categories according to the recall method corresponding to each of the selected index categories.

In one embodiment, the recall method includes at least one of entity database recall, terminology recall, and search engine recall. The entity database recall method is used to indicate the recall of category content related to the search term from the entity database corresponding to the index category. The entity database is a classification database pre-built by aggregating and classifying preset category content; category content belonging to the same index category is aggregated in the same entity database. The terminology recall method is used to indicate the recall of category content related to the search term from the terminology corresponding to the index category. The terminology is a terminology pre-built by aggregating and classifying preset category content; category content belonging to the same index category is aggregated in the same terminology.

In one embodiment, the index categories include at least one of the following: sub-application index category, public account index category, tool index category, multimedia content index category, and question and answer index category.

In one embodiment, as shown in FIG10, a search device is provided. This device can be a software module, a hardware module, or a combination of both, integrated into a computer device. Specifically, the device includes: a receiving module 1002, an index category filtering module 1004, an acquisition module 1006, and an aggregation suggestion module 1008, wherein:

The receiving module 1002 is used to receive the search terms entered on the query page.

The index category filtering module 1004 is used to filter multiple index categories that are related to the query intent of the search terms from the index categories supported by the search engine.

The acquisition module 1006 is used to acquire category content related to the search term under multiple filtered index categories.

The aggregation suggestion module 1008 is used to generate and return aggregation suggestion content corresponding to the search terms based on the filtered index categories and category content; the aggregation suggestion content is used to indicate whether to display the aggregation suggestion content on the query page.

In one embodiment, the aggregated prompt content further includes at least one of intent-related prompts associated with the search term and a search result viewing entry; wherein, the intent-related prompts are used to redirect to the search results page obtained by searching with the intent-related prompts as search criteria after being triggered; and the search results viewing entry is used to redirect to the search results page obtained by searching with the search term as search criteria after being triggered.

In one embodiment, the acquisition module 1006 is further configured to recall candidate category content related to the search term under the filtered plurality of index categories; when there are multiple candidate category contents under the same index category, the preset number of candidate category contents are selected from the multiple candidate category contents under the same index category in descending order of text relevance between each candidate category content and the search term, and used as category content for generating aggregated prompt content.

In one embodiment, the acquisition module 1006 is further configured to determine the recall method corresponding to each of the selected index categories based on the search term; for each selected index category, the recall method corresponding to the index category is used to recall the category content related to the search term under the index category.

In one embodiment, the aggregation suggestion module 1008 is further configured to sort the index categories in descending order of category intent intensity corresponding to the filtered index categories, or sort the category content in descending order of relevance between the category content and the search term; and generate and return aggregate suggestion content corresponding to the search term based on the sorted index categories and the category content under the index categories.

In one embodiment, the aggregation suggestion module 1008 is further configured to obtain a user profile of the user who inputs the search term; determine the matching degree between the category content and the user profile; sort the category content in descending order of the matching degree; and generate and return aggregate suggestion content corresponding to the search term based on the sorted index categories and the category content under the index categories.

Specific limitations regarding the search device can be found in the limitations of the search method above, and will not be repeated here. Each module in the aforementioned search device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in the computer device in hardware form, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in Figure 11. The computer device includes a processor, memory, communication interface, display screen, and input device connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for wired or wireless communication with external terminals. Wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a search method. The display screen of the computer device may be a liquid crystal display (LCD) or an e-ink display. The input device of the computer device may be a touch layer covering the display screen, or buttons, a trackball, or a touchpad located on the casing of the computer device, or an external keyboard, touchpad, or mouse, etc.

In one embodiment, a computer device, which may be a server, is provided, and its internal structure diagram is shown in Figure 12. The computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores search data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements a search method.

Those skilled in the art will understand that the structures shown in Figures 11 and 12 are merely block diagrams of some structures related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements.

In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims (20)

1. A search method, characterized in that the method comprises: The query page displays an input area, which includes a search box. The query page is for users to enter search terms to trigger a search engine to search for results on an internet platform. When a search term is detected in the search box, a dropdown menu is displayed below the search box on the query page. The dropdown menu sorts and displays the category content based on at least one of the following: the matching degree between the category content and the user profile; the category intent strength corresponding to the index category to which the category content belongs; and the relevance between the category content and the search term. The corresponding index category is also displayed for each category content. The user profile refers to the user profile of the user who entered the search term. The category content is included in the aggregated suggestion content. The aggregated suggestion content is set in the content aggregation card. The aggregated suggestion content refers to content composed of multiple suggestion items. The aggregated suggestion content further includes multiple index categories and intent-related suggestion words associated with the search term. The multiple index categories are those selected from the search intent-related categories supported by the search engine after the monitored search term is sent to the server, where the server performs category intent analysis on the search term. The category content is the category content related to the search term under the selected index categories, retrieved after the index categories have been selected. The index categories supported by the search engine refer to the categories that can provide query results when searching in the search engine. When a trigger operation targeting the intent-related prompt word is detected, the user is redirected to the first search results page, which displays search results obtained using the triggered intent-related prompt word as the search criteria. 2. The method according to claim 1, wherein the index categories include at least one of the following: sub-application index category, public account index category, tool index category, multimedia content index category, information index category, entertainment index category, or Q&A index category. 3. The method according to claim 1, characterized in that the method further comprises: When a trigger operation is detected on the category content displayed in the aggregated prompt, then Redirecting to the page corresponding to the triggered category content; wherein the page to which the redirection occurs corresponds to the index category to which the triggered category content belongs. 4. The method according to claim 1, wherein the intent-related prompt includes at least one of related keywords and extended keywords; the related keywords refer to words that include keywords in the search term; and the extended keywords are keywords that are not included in the keywords in the search term but are related to the query intent. 5. The method according to any one of claims 1 to 4, characterized in that the aggregated prompt content further includes a search result viewing entry; the method further includes: When a trigger operation targeting the search result viewing entry is detected, then Redirecting to the second search results page; The second search results page displays the search results obtained using the search term as the search criteria. 6. A search method, characterized in that the method comprises: Receive the search terms entered on the query page; the query page is a page where users enter search terms to trigger a search engine to search for results on an internet platform; The search terms are subjected to category intent analysis to filter multiple index categories related to the query intent of the search terms from the index categories supported by the search engine; the index categories supported by the search engine refer to the categories that can return the query content when searching in the search engine. Obtain the category content related to the search term under the filtered multiple index categories; The index categories are sorted in descending order of category intent intensity, and the category content is sorted in descending order of relevance between the category content and the search term, or in descending order of matching degree between the category content and the user profile. Based on the sorted index categories, the category content under each index category, and intent-related prompts associated with the search term, aggregated prompt content corresponding to the search term is generated and returned. The user profile refers to the user profile of the user who entered the search term. The aggregated prompt content is used to indicate the display of a dropdown menu below the search box on the query page, and to display a content aggregation card within the dropdown menu, so as to display the aggregated prompt content in the content aggregation card. The aggregated prompt content refers to content aggregated from multiple prompts. The index categories supported by the search engine refer to the categories that can provide query results when searching in the search engine. The intent-related prompts are used to redirect to the search results page obtained by searching using the intent-related prompts as search conditions after being triggered. 7. The method according to claim 6, wherein the aggregated prompt content further includes a search result viewing entry; The search results viewing entry is used to redirect to the search results page when triggered, based on the search terms. 8. The method according to claim 6, wherein obtaining the category content related to the search term under the filtered plurality of index categories includes: Under the selected multiple index categories, retrieve category content related to the search term; When there are multiple categories under the same index category, then Based on the text relevance between each category content and the search term, from high to low, a predetermined number of category contents are selected from multiple category contents under the same index category as the category contents used to generate aggregated suggestion content. 9. The method according to claim 6, wherein recalling category content related to the search term under the selected plurality of index categories includes: Based on the search terms, determine the recall method corresponding to each of the selected index categories; For each selected index category, retrieve the category content related to the search term under the index category according to the recall method corresponding to the index category. 10. A search device, characterized in that the device comprises: The display module is used to display the query page; the query page displays an input area; the input area includes a search box; the query page is a page for users to enter search terms to trigger a search query; A listening module is used to listen for search terms entered in the search box; The display module is further configured to, when the monitoring module detects a search term being entered in the input area, display a dropdown menu below the search box on the query page, and sort and display the category content in the dropdown menu according to at least one of the following: the matching degree between the category content and the user profile, the category intent strength corresponding to the index category to which the category content belongs, and the relevance between the category content and the search term, and display the corresponding index category corresponding to the category content; the user profile refers to the user profile of the user who entered the search term; the category content is included in the aggregated prompt content; the aggregated prompt content is set in the content aggregation card; the aggregated prompt content refers to content aggregated from multiple prompts; The aggregated suggestion content further includes multiple index categories and intent-related suggestion words associated with the search term. The multiple index categories are those selected from the search intent-related categories supported by the search engine after the monitored search term is sent to the server, where the server performs category intent analysis on the search term. The category content is the category content related to the search term under the selected index categories, retrieved after the index categories have been selected. The index categories supported by the search engine refer to the categories that can provide query results when searching in the search engine. The display module is also used to redirect to the first search results page when the monitoring module detects a trigger operation for the intent-related prompt word; wherein, the first search results page displays the search results obtained using the triggered intent-related prompt word as the search condition. 11. The apparatus according to claim 10, wherein the index categories include at least one of the following: sub-application index category, public account index category, tool index category, multimedia content index category, information index category, entertainment index category, or question and answer index category. 12. The apparatus according to claim 10, wherein the display module is further configured to, when the monitoring module detects a trigger operation on the category content displayed in the aggregated prompt content, jump to the page corresponding to the triggered category content; wherein the page to which the jump is made conforms to the index category to which the triggered category content belongs. 13. The apparatus according to claim 10, wherein the intent-related prompt includes at least one of related keywords and extended keywords; the related keywords refer to words that include keywords in the search term; and the extended keywords are keywords that are not included in the keywords in the search term but are related to the query intent. 14. The apparatus according to any one of claims 10 to 13, wherein the aggregated prompt content further includes a search result viewing entry; the display module is further configured to, when the monitoring module detects a trigger operation on the search result viewing entry, jump to a second search result page; wherein, the second search result page displays search results obtained using the search term as the search condition. 15. A search device, characterized in that the device comprises: The receiving module is used to receive the search terms entered on the query page; the query page is a page where users can enter search terms to trigger a search engine to search for results on an internet platform. The index category filtering module is used to filter multiple index categories related to the query intent of the search term from the index categories supported by the search engine; the index categories supported by the search engine refer to the categories that can provide query content when searching in the search engine. The acquisition module is used to acquire category content related to the search term under the filtered multiple index categories; The aggregation suggestion module is used to sort the index categories according to the category intent intensity from high to low, and to sort the category content according to the relevance between the category content and the search term from high to low, or according to the matching degree between the category content and the user profile from high to low. Based on the sorted index categories, the category content under the index categories, and intent-related suggestion words associated with the search term, it generates and returns aggregated suggestion content corresponding to the search term. The user profile refers to the user profile of the user who input the search term. The aggregated suggestion content is used to indicate the display of a drop-down box below the search box on the query page, and to display a content aggregation card in the drop-down box to display the aggregated suggestion content. The aggregated suggestion content refers to content aggregated from multiple suggestion items. The index categories supported by the search engine refer to the categories that can provide query results when searching in the search engine. The intent-related suggestion words, when triggered, redirect to the search results page obtained by searching using the intent-related suggestion words as search conditions. 16. The apparatus according to claim 15, wherein the aggregated prompt content further includes a search result viewing entry; the search result viewing entry is used to redirect to a search results page obtained by performing a search using the search term as the search condition after being triggered. 17. The apparatus according to claim 15, wherein the acquisition module is further configured to recall category content related to the search term under the filtered plurality of index categories; when there are multiple category contents under the same index category, the preset number of category contents are selected from the multiple category contents under the same index category in descending order of text relevance between each category content and the search term, and used as category contents for generating aggregated prompt content. 18. The apparatus according to claim 15, wherein the acquisition module is further configured to determine, based on the search term, the recall method corresponding to each of the selected index categories; and for each selected index category, to recall the category content related to the search term under the index category according to the recall method corresponding to the index category. 19. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method according to any one of claims 1 to 9. 20. A computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 9.