CN110134845A - Project public opinion monitoring method, device, computer equipment and storage medium - Google Patents

Abstract

Embodiments of the present application provide a project public opinion monitoring method, device, computer equipment, and computer-readable storage medium. The embodiment of the present application belongs to the field of data display technology. When realizing project public opinion monitoring, after obtaining the identification information of the target project, the data source website list of the target project is obtained through network search according to the identification information, and the data source website list is obtained from the data according to the identification information. The corpus of the target project is crawled from the data source website contained in the source website list, and the corpus about the target project is obtained, and then the corpus is parsed through natural language processing to identify the subject name and public opinion features contained in the corpus, and the subject name and The public opinion features are imported into the graph database to construct the public opinion relationship map of the target project, and the public opinion relationship map of the target project is visually displayed, so as to monitor the public opinion of the project from the perspective of subdivision of the target unit, thereby improving the detailed information within the target unit. The efficiency of public opinion monitoring from different angles.

Description

Project public opinion monitoring method, device, computer equipment and storage medium

technical field

The present application relates to the technical field of data display, and in particular to a project public opinion monitoring method, device, computer equipment, and computer-readable storage medium.

Background technique

Corporate public opinion information is currently involved in all corporate relationship projects, while traditional technology uses a targeted method of collecting data from designated websites, such as financial websites. However, the information collected in this way reflects the overall public opinion of an enterprise. If public opinion on a certain aspect of an enterprise is needed, such as public opinion on a product, an investment, or an advertising campaign of the enterprise, it is necessary to collect information from a large number of sources. The overall public opinion information data is screened, and the screened content is not accurate enough, which leads to the problem of low efficiency in obtaining public opinion in this area.

Contents of the invention

The embodiment of the present application provides a project public opinion monitoring method, device, computer equipment, and computer-readable storage medium, which can solve the problem of low efficiency of target project public opinion monitoring in the traditional technology.

In the first aspect, an embodiment of the present application provides a project public opinion monitoring method, the method comprising: obtaining the identification information of the target project in a preset manner; obtaining the data of the target project through a network search according to the identification information Source website list; crawl the corpus of the target item from the data source website contained in the data source website list according to the identification information; analyze the corpus by natural language processing to identify the subject name contained in the corpus and public opinion characteristics; importing the subject name and the public opinion characteristics into a graph database to construct a public opinion relationship graph of the target project; displaying the public opinion relationship graph.

In the second aspect, the embodiment of the present application also provides a project public opinion monitoring device, including: a first acquisition unit, configured to acquire the identification information of the target item in a preset manner; a second acquisition unit, configured to Obtaining a list of data source websites of the target item through network search; a crawling unit configured to crawl the corpus of the target item from the data source websites included in the list of data source websites according to the identification information; The identification unit is used to analyze the corpus through natural language processing to identify the subject name and public opinion features contained in the corpus; the construction unit is used to import the subject name and the public opinion features into the graph database to construct the target A public opinion relationship map of the project; a display unit configured to display the public opinion relationship map.

In the third aspect, the embodiment of the present application also provides a computer device, which includes a memory and a processor, the memory stores a computer program, and the processor implements the project public opinion monitoring method when executing the computer program.

In the fourth aspect, the embodiment of the present application also provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor executes the project public opinion monitoring method.

Embodiments of the present application provide a project public opinion monitoring method, device, computer equipment, and computer-readable storage medium. When the embodiment of the present application implements project public opinion monitoring, after obtaining the identification information of the target project, obtain the data source website list of the target project through network search according to the identification information, and obtain the website list of the data source of the target project according to the identification information from the data source The corpus of the target item is crawled from the data source website contained in the website list, so as to obtain a more comprehensive corpus about the target item, and then analyze the corpus through natural language processing to identify the subject name and Public opinion features, importing the subject name and the public opinion features into the graph database to construct the public opinion relationship map of the target project, and visually display the public opinion relationship map of the target project, so as to realize the target unit from the perspective of subdivision. Public opinion monitoring, so as to improve the efficiency of public opinion monitoring from the subdivision angle within the target unit.

Description of drawings

In order to illustrate the technical solutions of the embodiments of the present application more clearly, the drawings that need to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are some embodiments of the present application. Ordinary technicians can also obtain other drawings based on these drawings on the premise of not paying creative work.

Fig. 1 is a schematic diagram of the application scenario of the project public opinion monitoring method provided by the embodiment of the present application;

Fig. 2 is a schematic flow diagram of the project public opinion monitoring method provided by the embodiment of the present application;

Fig. 3 is another schematic flow chart of the project public opinion monitoring method provided by the embodiment of the present application;

Fig. 4 is a schematic diagram of a sub-flow of the project public opinion monitoring method provided by the embodiment of the present application;

Fig. 5 is a schematic diagram of another sub-flow of the project public opinion monitoring method provided by the embodiment of the present application;

6 is a schematic diagram of the third sub-flow of the project public opinion monitoring method provided by the embodiment of the present application;

FIG. 7 is a schematic block diagram of a project public opinion monitoring device provided in an embodiment of the present application;

Fig. 8 is another schematic block diagram of the project public opinion monitoring device provided by the embodiment of the present application; and

Fig. 9 is a schematic block diagram of a computer device provided by an embodiment of the present application.

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of them. Based on the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the scope of protection of this application.

It should be understood that when used in this specification and the appended claims, the terms "comprising" and "comprises" indicate the presence of described features, integers, steps, operations, elements and/or components, but do not exclude one or Presence or addition of multiple other features, integers, steps, operations, elements, components and/or collections thereof.

It should also be understood that the terminology used in the specification of this application is for the purpose of describing particular embodiments only and is not intended to limit the application. As used in this specification and the appended claims, the singular forms "a", "an" and "the" are intended to include plural referents unless the context clearly dictates otherwise.

It should also be further understood that the term "and/or" used in the description of the present application and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations .

Please refer to FIG. 1 . FIG. 1 is a schematic diagram of an application scenario of the project public opinion monitoring method provided by the embodiment of the present application. The project public opinion monitoring method provided in the embodiment of the present application can be applied to the terminal shown in Figure 1, and the steps of the project public opinion monitoring method are realized through the software installed on the terminal, wherein the terminal can be a notebook computer or a tablet computer Or electronic equipment such as desktop computers. The specific implementation process of the project public opinion monitoring method provided by the embodiment of the present application is as follows: the terminal obtains the identification information of the target project in a preset manner; obtains the data source website list of the target project through a network search according to the identification information; The identification information crawls the corpus of the target project from the data source website included in the data source website list; parses the corpus through natural language processing to identify the subject name and public opinion features contained in the corpus; The subject name and the public opinion features are imported into a graph database to construct a public opinion relationship graph of the target item; and the public opinion relationship graph is displayed.

It should be noted that in FIG. 1 only a desktop computer is shown as a terminal. In the actual operation process, the type of the terminal is not limited to that shown in FIG. The application scenarios of the project public opinion monitoring method above are only used to illustrate the technical solution of the present application, and are not intended to limit the technical solution of the present application.

Fig. 2 is a schematic flow chart of a project public opinion monitoring method provided by an embodiment of the present application. The project public opinion monitoring method is applied to the terminal shown in FIG. 1 to complete all or part of the functions of the project public opinion monitoring method.

Please refer to FIG. 2 . FIG. 2 is a schematic flowchart of a project public opinion monitoring method provided by an embodiment of the present application. As shown in Figure 2, the method includes the following steps S210-S260:

S210. Obtain the identification information of the target item in a preset manner.

Wherein, the preset method refers to a method of manually inputting through an input device or a method of obtaining vocabulary through natural language processing of the corpus of the target item and screening the obtained vocabulary .

The target project refers to the project that the enterprise or other organizations determine to monitor public opinion. For example, the target project of an enterprise refers to the content of a certain aspect of the enterprise, such as a product of the enterprise, a marketing, an investment, an event, etc. Content is the content of a segmented aspect of an enterprise.

The identification information of the target item refers to the identification information for the main content of the target item, which is a description of the key content of the target item, for example, for a product, the name and various attributes of the product, such as a mobile phone, etc. .

Specifically, the identification information of the target item can receive input information through an input device, thereby obtaining the identification information of the item to monitor public opinion on the item, for example, obtaining the brand name, model, processor and battery life of a mobile phone and other performance descriptions. In addition, the identification information of the target project can also be obtained through natural language processing. The initial corpus is crawled from the corpus source of the target project, and the initial corpus is subjected to natural language processing such as word segmentation and screening, and the target project is currently concerned. According to the identification information of the target project, more comprehensive corpus is crawled for the target project according to the identification information, and more corpus is used to accurately monitor the public opinion of the target project. For example, the terminal crawls the data of an enterprise for a predetermined period, screens out the current hot target items of the company from the corpus through natural language processing, and obtains the identification information of the target items, that is, screens out the hot items that the company is currently concerned about. Public opinion, according to the identification information, crawls the data of the target project involved in the identification information, further analyzes the data, and obtains the public opinion of the target project for reference by enterprises.

S220. Obtain a list of data source websites of the target item by searching the Internet according to the identification information.

Specifically, the terminal obtains the identification information of the target item, and according to the identification information of the target item, obtains the corpus source involved in the target item through searching, and the corpus source is the data source website of the target item. source website, it is a list of data source websites. For example, in order to understand the public opinion on a product project of the enterprise, that is, to understand the external evaluation of the product, an enterprise needs to monitor the public opinion of the product, and can obtain the keywords of the product, such as the brand name and model of the product etc., according to the keyword, obtain the data source involved in the product, of course, the data source can also be a manually input URL, according to the input URL and keyword, obtain the corpus involved in forming the public opinion of the product, the The corpus includes news, product introductions, product reviews and comments, etc., and can be described in the form of articles or sentences and exists in various types of websites. For example, in order to realize the public opinion monitoring of a mobile phone product and improve the marketing and design of mobile phone products, etc., an enterprise needs to obtain keywords of mobile phone products, such as mobile phone name, mobile phone model and other keywords. According to the keywords of mobile phone products, obtain Contains the data source of the mobile phone product, and obtains the corpus of public opinion for the mobile phone product from the data source.

S230. Crawl the corpus of the target item from the data source websites included in the data source website list according to the identification information.

Among them, crawling refers to crawling through crawlers, crawlers refer to web crawlers, and web crawlers are also called web spiders, web robots, or web page chasers, etc., and are a program that automatically grabs information on the World Wide Web according to certain rules. or script.

Specifically, it is necessary to implement public opinion monitoring on the target project, crawl the relevant corpus of the target project on the Internet by building a crawler system, construct a public opinion relationship map of the target project through the analysis of the corpus, and obtain the public opinion of the target project through the public opinion relationship map. In order to realize the public opinion monitoring of the target project. Since a web crawler is a program that automatically extracts web pages, and crawls the corpus containing the target item in the data source website, only the data related to the target item can be crawled through the crawler program, and the identification of the target item can be obtained by a preset method Information, after obtaining the data source website list of the target project through the identification information, the crawler system can obtain rich corpus of the target project in the data source website through crawling according to the data source website list of the target project.

Furthermore, the data contained in the corpus can also be screened, and the public opinion of a certain aspect of the target project public opinion can be obtained according to the screened data, for example, the camera evaluation for a certain mobile phone product, battery life, processor or system optimization. Inferior data are screened to form corresponding public opinion.

S240. Analyze the corpus by using natural language processing to identify subject names and public opinion features included in the corpus.

Among them, the target project refers to the overall content of the enterprise or other organizations to determine and monitor public opinion. For example, the target project of an enterprise can be a product, a marketing, an investment, an event, etc. of the enterprise. A subdivision of the content of the whole. The subject name refers to the object contained in the corpus, which is the name of the target project and the names of each component in the target project. For example, if a target project is mobile phone A, the subject name includes the title of mobile phone A and the composition of the mobile phone. The name of each component or component of A, because the corpus contains the name of mobile phone A, the name of the display screen B and the name of camera C that make up mobile phone A, in the process of recognition, the computer equipment cannot distinguish between the names of the various subjects For example, the display screen B and the camera C belong to the mobile phone A, but the name of the mobile phone A contained in the corpus can be identified, and the name of the display screen B and the name of the camera C that make up the mobile phone A are the subject names.

The public opinion feature refers to the keywords of the public opinion of the target project. It is a feature description for evaluating the target project. It is used to describe the attributes of the subject name corresponding to the subject in the target project and the relationship between subjects. For example, if the target project is a mobile phone, the target project The subject name contained in the subject name is the name of the mobile phone, the names of various parts or components in the mobile phone, such as the name of the display screen and the name of the camera that make up the mobile phone. The public opinion feature is the evaluation and description of the subject corresponding to the subject name. For example, the configuration of the mobile phone is high. , a large display screen, or a camera with a good photo effect, etc., which reflect the characteristics of the mobile phone. It should be noted that the subject name refers to the identification of the subject, and the subject name can be expressed in different forms. For example, for mobile phones, in addition to the mobile phone brand as the mobile phone name, the specific model or code name of the mobile phone brand can also be used as the mobile phone name.

Specifically, parsing the corpus through natural language processing refers to dividing the corpus according to the sentence separator to obtain a sentence data set, constructing a named entity model based on the corpus, and identifying the sentence through the named entity model The name of the subject contained in the data set, the part of speech analysis and the retrieval of the target relationship are performed on the corpus to obtain the public opinion characteristics of the target item. For example, through natural language processing technology, analyze the acquired corpus, identify the mobile phone name information and related feature descriptions, and provide an important data source for the public opinion of the target project. Among them, named entity recognition, English is Named Entity Recognition, NER for short, also known as "proper name recognition", which refers to the identification of entities with specific meaning in text, mainly including names of people, places, institutions, proper nouns, etc. Generally speaking, the task of named entity recognition is to identify three major categories (entity, time, and number) and seven subcategories (person name, organization name, place name, time, date, currency, and percentage) in the text to be processed. Entities, Chinese named entity models include CRF models and word-based BiLSTM-CRF models. Through the obtained detailed and comprehensive data sources, the public opinion information about the target project is obtained through natural language processing methods, and then the public opinion information is imported into the graph database to improve the data of nodes and node attributes. For example, through the named entity model, the sentence corpus related to the target project is identified, and after the corpus is segmented, the part of speech analysis and feature word analysis are performed on the words, such as nouns, verbs, adjectives and the relationship between these words, so as to extract the corpus The public opinion information of the target project in .

Furthermore, artificial intelligence model training and automatic learning technology can also be used to improve the accuracy of subject name recognition and public opinion feature recognition in target projects. Specifically, through natural language processing technology, a large amount of corpus is used for word segmentation processing, and the acquired word segmentation is screened. At this time, artificial intelligence model training and automatic learning technology are used to improve the accuracy of subject name recognition and public opinion feature recognition. For example, the nouns and verbs contained in the obtained word segmentation are screened out through the artificial intelligence model and automatic learning technology, and the pre-preset digits of the nouns and verbs are screened out in order from high to low, and the nouns are used as the target The subject of the project, the verb is used as the description of the relationship between subjects, and the entity model is established.

S250. Import the subject name and the public opinion features into a graph database to construct a public opinion relationship graph of the target item.

Among them, graph database, also known as graph database, is Graph Database in English. Graph database is a type of NoSQL database. It applies graph theory to store relationship information between entities. Common graph databases include Neo4j, FlockDB, and AllegroGrap. In a graph database, there are two main components of the database, the node set and the relationship between the connecting nodes. The node set is a collection of a series of nodes in the graph. In the graph database, each node still has The label is also the set of nodes it belongs to, and records a series of attributes describing the characteristics of the node. In addition, the nodes can also be connected through relationships.

Specifically, import the subject name and public opinion features of the target project identified through natural language processing and analysis of the corpus into the graph database, and improve the data of the nodes and connection node relationships of the graph database, wherein the nodes correspond to the subject name and public opinion features, while describing the relationship between nodes. When designing a graph database, a node set is composed of multiple nodes, and the nodes are associated through relationships. The node set in the graph, the nodes, and the relationship between nodes are distinguished. When importing data, the graph database automatically recognizes the data in the imported data. The node data and the relationship data are assigned to the corresponding positions of the graph database respectively. In this example, after importing the subject name and the public opinion features into the graph database, the public opinion relationship map of the target project can be automatically constructed. For example, if the target project is a mobile phone, the subject name contained in the target project is The name of the mobile phone, the name of each component or component in the mobile phone, such as the name of the display screen and the name of the camera that make up the mobile phone. The public opinion feature is the evaluation and description of the subject corresponding to the subject name. For example, the large display screen of the mobile phone is a reflection of the mobile phone features. Description, the three nodes correspond to "mobile phone", "display screen" and "big", and at the same time through the "containment" relationship between "mobile phone" and "display screen", and the description between "display screen" and "big" The relationship connects the three nodes in turn to form a public opinion relationship graph. In the embodiment of the present application, the dynamic public opinion data of the target item is stored by means of the public opinion relationship graph of the target item, so that the public opinion of the target item can be better visualized and extracted.

Further, the public opinion relationship graph of the target project includes the name of the target project, the names of the sub-projects of the target project, and the public opinion features corresponding to the names of the sub-projects.

Wherein, a sub-item of the target item refers to a component of the target item. For example, if the target item is a mobile phone, then components such as the display screen, camera, battery and central processing unit of the mobile phone are all sub-items of the mobile phone, and the model corresponding to the sub-item is the name of the sub-item.

Specifically, the elements in the public opinion relationship graph of the target item include subject names and public opinion features in the public opinion relationship graph of the target item. For example, the public opinion relationship graph of the target project includes subject names such as nouns and feature descriptions such as adjectives. Public opinion features refer to keywords of public opinion, for example, clear photos and long battery life. Through the public opinion relationship map of the target project, the dynamic public opinion data of the target project is stored to achieve better visualization and extraction of public opinion data. By constructing the map data of the target project, a news corpus related to the target project is built. In the target project Before graph data visualization, it is also necessary to sort the public opinion data of the target project by time, and list the recent news data of the target project that ranks relatively high. The relevant objects of the node can obtain the relevant public opinion of A product, etc. In addition, in-depth analysis of specific fields related to the target project can also be carried out. For example, if the target project is a product, the supplier relationship of the product can also be analyzed, and the public opinion data related to the target product needs to be obtained from the supplier attributes. , and classify and deduplicate the data, and present it to users for public opinion monitoring. For example, the evaluation of the CPU in mobile phone products can affect the public opinion of the product, such as the heating problem of the CPU Snapdragon 820 in smartphones , has a greater impact on mobile phones using the Snapdragon 820. Various praises for the performance advantages of the Snapdragon 835 processor have brought various praises for the performance of the Snapdragon 835 mobile phone, such as fast speed and power saving. characteristics of public opinion.

S260. Display the public opinion relationship graph.

Specifically, the terminal displays the constructed public opinion relationship map of the target project and provides it to the user so that the user can monitor the public opinion of the target project according to the public opinion relationship map of the target project, so that the target project monitors can monitor the public opinion according to the The public opinion relationship map of the target project obtains the conclusion of the public opinion of the target project, realizes the monitoring of the public opinion of the target project, and handles the public opinion of the target project correspondingly, for example, can obtain the positive information and negative information of the public opinion of the target project, and obtain the events in the public opinion of the target project Evaluation information and channel evaluation information for corresponding public relations measures.

Furthermore, it is also possible to sort the positive and negative public opinion conclusions of the obtained target project public opinion according to different mechanisms, make full use of the positive public opinion to maximize the benefits, and take corresponding measures against the negative public opinion to eliminate the negative impact. For example, a certain product appears screen problems or battery problems.

When the embodiment of the present application implements project public opinion monitoring, after obtaining the identification information of the target project, obtain the data source website list of the target project through network search according to the identification information, and obtain the website list of the data source of the target project according to the identification information from the data source The corpus of the target item is crawled from the data source website contained in the website list, so as to obtain a more comprehensive corpus about the target item, and then analyze the corpus through natural language processing to identify the subject name and Public opinion features, importing the subject name and the public opinion features into the graph database to construct the public opinion relationship map of the target project, and visually display the public opinion relationship map of the target project, so as to realize the target unit from the perspective of subdivision. Public opinion monitoring, increasing the specific public opinion of the target project, can better realize the public opinion monitoring for a specific project in the target unit, so as to judge the pros and cons of the project. In this way, the efficiency of public opinion monitoring from the subdivision angle within the target unit can be improved.

Please refer to FIG. 3 . FIG. 3 is another schematic flowchart of the project public opinion monitoring method provided by the embodiment of the present application. In this embodiment, after the step of obtaining the list of data source websites of the target item by searching the Internet according to the identification information, it further includes:

S221. Update the data source website list by crawling.

Specifically, build a crawler strategy that automatically adds data sources, and obtain more comprehensive data sources of target projects from the Internet through deep crawling. The crawler strategy that can automatically increase the data source means that after the crawler receives the initialized data source website, it can automatically expand more data source websites to increase the source of corpus according to the obtained data source website, so as to obtain a more comprehensive corpus of the target project . In this embodiment, the crawler strategy that can automatically increase the data source means that the crawler digs out a new data source associated with the obtained data source website by crawling according to the type of the obtained data source website and the structure characteristics of the URL Websites, for example, have the same suffix as the obtained data source website, or belong to the same type as the obtained data source website, for example, both belong to financial websites, etc., so as to expand from one financial website to other financial websites. There may be corpus interpreted from different angles for the same target item. Because the websites related to each other, especially when facing the hot issues of the target project, they will interpret and report the target project from different angles, so as to continuously improve the websites in the data source website and enrich the information in the data source website. Data sources, to increase data sources and ensure the basis of data volume. Obtain the relevant corpus of the target project through the data source website, and obtain comprehensive and rich corpus of the target project through rich data sources. Further, the crawler strategy of automatically adding data sources can be used to build a real-time distributed crawler system, and improve the efficiency of crawling data through the distributed crawler system. Specifically, the server obtains the initial data source website list of the target project, classifies the initial data source website list according to preset conditions to obtain different types of data source website lists, and encapsulates the different types of data source website lists into corresponding Docker container, deploy different Docker containers on different servers, start the Docker container so that the Docker container can obtain the new data source website by crawling, and add the new data source website to the corresponding initial List of data source sites to update the data source sites for said target project. For example, constructing a crawler strategy for automatically adding data sources is a real-time distributed crawler system. The crawler system can distinguish the types of different websites according to the input list, such as the identification of the website URL in the input list. According to the type of website , assign the list to each server, realize distributed data crawling and data storage, and improve the efficiency of crawling data.

Please refer to FIG. 4 . FIG. 4 is a schematic diagram of a sub-flow of the project public opinion monitoring method provided by the embodiment of the present application. As shown in Figure 4, in this embodiment, the step of updating the list of data source websites by crawling includes:

S2210. Obtain an initial data source website list of the target item;

S2211. Classify the initial data source website list according to preset conditions to obtain different types of data source website lists;

S2212. Encapsulating the different types of data source website lists into corresponding Docker containers;

S2213. Start the Docker container so that the Docker container obtains the new data source website by crawling;

S2214. Add the new data source website according to type to the corresponding classified data source website list to update the data source website of the target project.

Among them, the preset conditions include conditions such as website address or data source, and the website address refers to classifying according to the Uniform Resource Locator (Uniform Resource Locator in English, abbreviated as URL) of the website. Since different websites have different anti-crawler strategies, As a result, the data structure of the web pages in the website is different, and different crawling strategies are required for different websites. For example, the news of Sina.com is easier to crawl, and it can be directly parsed by BeautifulSoup and directly crawled. The title of NetEase news and The content is loaded asynchronously using JS. Simply downloading the source code of the web page does not have the title and content. You can find the required content in the JS of the Network. You can use regular expressions to obtain the title and its link we need. Toutiao's The news is different from the previous two. Its title and link are encapsulated into the Json file, but the URL parameters of the Json file are changed through a JS random algorithm. It is necessary to simulate the parameters of the Json file, otherwise the specific details of the Json file cannot be found. URL, website sources include financial websites, news websites or forums, etc.

Specifically, the initial data source website list of the configured target project is obtained, and the crawler system automatically classifies the initial data source website list according to the preset conditions of the initial data source website list to obtain different types of data source website lists, For example, the data source website is divided into different types according to the website identifier, and then the list of different types of data source websites is packaged into a corresponding Docker container, and the Docker container is deployed on a different server, and the Docker container is started so that all The Docker container obtains rich new data source websites by crawling, and adds the new data source websites to the corresponding initial data source website list to update the data source website of the target project, thereby continuously improving the data source website of the target project . Specifically, the following sub-steps are included:

First, an initial website list is obtained. The list can be manually configured, that is, an initial data source website is provided manually, or it can be a website list searched based on identification information.

Secondly, by encapsulating the written crawler code into a Docker container, the code includes the part of extracting the URL of the website, as well as the code for matching the URL and the corresponding crawler program, so that the URL automatically corresponds to the crawler program, and through the corresponding The crawler program crawls the website corresponding to the URL. Among them, it is necessary to construct an index relationship between URLs and crawlers, and prepare web crawlers for all URL types in advance, so that different types of URL crawlers correspond to different crawler programs.

Third, start the container Docker1, classify and divide the total input list through the crawler code, save the list of data sources of the same type, and form a list to be crawled, waiting to be crawled. Among them, by starting the URL classification and segmentation code, the input website URL list is classified according to the URL type, and the website URL list is classified, and then, the list segmentation code is started, and different data source lists are divided into several lists. Corresponds to Docker containers on different machines.

Fourth, start the container Docker2, through the obtained list of data sources, by matching the crawler program corresponding to the URL, for example, X website, corresponding to the crawling and parsing code of the X website, and then uploading the X website to crawl, external The network is accessed, the corresponding data is fetched separately, and the data is returned to the database.

Further, the crawler program digs out new URLs according to the obtained URLs, that is, the crawler program digs out new URLs by starting the URL, and stores the new URLs in the URL list to be crawled to complete the URL list. At the same time, you can also check whether there is an error reported in the process of crawling data. If there is an error reported, the crawling process for this website ends.

To classify URLs, you can use preset URL regular expressions. Each type of URL list has a corresponding regular expression. By judging whether the returned result is empty, it is determined whether it is a URL of this type. The judging process is as follows: if the returned result is not empty, it is judged as a URL of this type; if the result of judging is empty, it is judged as a URL of not this type.

Fifth, until all the Docker2 website lists to be crawled are empty, stop the operation. In order to improve the list of data source websites, the above steps can be repeated according to the obtained list of data source websites in a regular or irregular manner, so as to update the list of data source websites.

Please refer to FIG. 5 . FIG. 5 is a schematic diagram of another sub-flow of the project public opinion monitoring method provided by the embodiment of the present application. As shown in Figure 5, in this embodiment, the step of analyzing the corpus to identify the subject name and public opinion features contained in the corpus through natural language processing includes:

S2400. Divide the corpus according to sentence separators to obtain sentence data sets.

Wherein, the sentence separator includes sentence punctuation marks and decomposition words, and the sentence punctuation marks include punctuation marks such as ".", "?", ";" and "!", and the decomposition words include "of", "and" , "中", "We" and "Based on" and other preset words or words that can be used as sentence separators.

Specifically, the corpus crawled by the crawler system is separated according to the sentence separator to obtain a sentence data set, so as to filter out sentences containing names from the sentence data set.

S2401. Construct a named entity model according to the corpus.

Among them, the named entity, English is Named Entity, the so-called named entity is the name of the person, the name of the organization, the place name and all other entities identified by the name, and the wider entity also includes numbers, dates, currencies, addresses and so on.

Specifically, in the construction of the named entity model, the named entity is marked through the acquired corpus content, and the named entity recognition model is constructed through the CRF model to identify the subject name of the target project. Among them, the CRF model, CRF, Conditional Random Field in English, conditional random field, is one of the commonly used algorithms in the field of natural language processing in recent years, based on statistical models, CRF is essentially a Markov chain of hidden variables and Conditional probability of observable states to hidden variables.

S2402. Using the named entity model, identify subject names included in the sentence dataset.

Among them, named entity recognition, English is Named Entity Recognition, NER for short, also known as "proper name recognition", which refers to the identification of entities with specific meaning in text, mainly including names of people, places, institutions, proper nouns, etc.

Specifically, after the named entity model is constructed, the obtained sentence data set is processed through the named entity model, and the topic names contained in the sentence data set can be automatically identified through the named entity model. For example, the named entity is marked through the content of the corpus, and the named entity recognition model is constructed through the CRF model to identify the name of the company entity. Through the named entity model, the sentence corpus in the relevant information of the target project is identified, and the word part of speech is analyzed and the key relationship of the project is retrieved. If there are core keywords, such as mobile phone products, such as long-lasting battery life and convenient system operation, etc., save the relevant information as a specific attribute of the target project. At the same time, the specific attribute can also carry the current date and time to enrich the public opinion data of the public opinion relationship graph of the target project.

S2403. Perform part-of-speech analysis and target relationship retrieval on the corpus to obtain public opinion features of the target item.

Among them, the part of speech refers to the characteristics of words as the basis for dividing part of speech, such as verbs, nouns and other parts of speech. The target relationship refers to the relationship between subjects involved in the target item included in the corpus, for example, the affiliation relationship between the mobile phone and components of the mobile phone, for example, the inclusion relationship between the mobile phone and the camera of the mobile phone, etc.

Specifically, performing part-of-speech analysis and subject relationship identification on the corpus includes the following processes

First, word segmentation is performed on the corpus. The word segmentation operation for the sentence type can use stuttering word segmentation. Among them, stammering word segmentation is one of the word segmentation tools in Python. There are many word segmentation tools in Python, including Pangu word segmentation, Yaha word segmentation, Jieba word segmentation, Tsinghua THULAC, etc.

Second, the core relationship is extracted. Specifically, the action of the verb is extracted, and the keyword list is matched. If the verb vocabulary is in the keyword, it is identified as the core relationship, and the noun object behind the verb is obtained, which is the object of the named relationship, and the object in front of the verb is obtained The noun object is the subject of the named relationship, and the subject of the named relationship is also the target. The acquired naming relationship subjects, naming relationship objects, and the relationship between naming relationship subjects and naming relationship objects are used as public opinion features, and the extracted core relationship involves subject names and characteristic data reflecting attributes are stored in the graph database.

Further, the step of constructing a named entity model according to the corpus includes:

1), carry out word segmentation to described corpus to obtain word segmentation result;

2), extracting the feature data in the word segmentation result by a preset feature template;

3) Training a preset conditional random field model based on the feature data to construct a named entity model.

Specifically, constructing a named entity model through the acquired corpus includes the following steps:

First, the named entity training corpus is obtained, which mainly comes from the corpus of the target item obtained by the crawler system through crawling.

Second, the corpus is preprocessed. Mainly use stuttering word segmentation and remove stop words and meaningless words to obtain word segmentation results.

Third, perform feature extraction. Feature extraction is performed through feature templates composed of regular expressions, and the acquired features include words, parts of speech, boundary words, and named entity feature words.

Fourth, create and train models based on conditional random fields. The conditional random field is also the CRF model. The CRF model is trained through the training data, the parameters of the CRF model are obtained, and the trained CRF model is saved.

Fifth, through the evaluation of the test data, and retain the model with a high recognition rate that finally meets the requirements, to obtain the constructed named entity model.

Please refer to FIG. 6, which is a schematic diagram of the third sub-flow of the project public opinion monitoring method provided by the embodiment of the present application. In this embodiment, the step of parsing the corpus through natural language processing to identify subject names and public opinion features contained in the corpus includes:

S2500. Segment the corpus to obtain a vocabulary list of the corpus.

Specifically, for the preprocessing of the corpus, stammer word segmentation is mainly used and stop words and meaningless words are removed to obtain a vocabulary list.

S2501. Use the first regular expression to extract the core relationship in the vocabulary list to obtain public opinion features;

S2502. Use the second regular expression to extract the named entities involved in the core relationship in the vocabulary list to obtain the subject name.

Specifically, regular expressions are used to extract core relations. Specifically, the action of the verb is extracted through regular expressions, and the keyword list is matched. If the verb vocabulary is in the keyword, it is identified as the core relationship, and the noun object behind the verb is obtained, which is the object of the named relationship. The noun object in front of the verb is the subject of the named relation, and the subject of the named relation is also the target. The obtained naming relationship subject, naming relationship object, and the relationship between the naming relationship subject and the naming relationship object are used as public opinion features. Among them, regular expressions, also known as regular expressions, are Regular Expressions in English, and are often abbreviated as Regex, Regexp or RE in code. Regular expressions are usually used to retrieve and replace text that conforms to a certain pattern (rule).

Then, import the core relationship as the public opinion feature and the named entity as the subject name into the graph database to construct the public opinion relationship graph of the target project. When designing a graph database, distinguish between node sets, nodes, and relationships in the graph. When importing data, the graph database automatically recognizes the node data and relationship data in the imported data, and assigns the node data and relationship data to the to the corresponding location in the graph database. In this example, after the core relationship and the named entity are imported into the graph database, the public opinion relationship graph of the target item can be automatically constructed. For example, the core relationship is used to construct the relationship between mobile phones and mobile phone components, and the feature description of components is imported into the attributes of component nodes. Among them, graph database, also known as graph database, is Graph Database in English. Graph database is a type of NoSQL database. It applies graph theory to store relationship information between entities. Common graph databases include Neo4j, FlockDB, and AllegroGrap.

Please continue to refer to Fig. 3, as shown in Fig. 3, in this embodiment, after the step of displaying the public opinion relationship map, it also includes:

S270. Combine the elements in the public opinion relationship graph in a preset order to describe the public opinion of the target item in text form.

Further, the step of combining elements in the public opinion relationship graph in a preset order to describe the public opinion of the target item in text includes:

Combining the elements in the public opinion relationship graph in a preset order to describe the positive public opinion information, negative public opinion information, event evaluation information and channel evaluation information of the target project in text form.

Specifically, not only displaying the public opinion of the target project in the form of the public opinion relationship map of the target project to realize the public opinion monitoring of the target project, but at the same time, by combining the display form of the text, giving the public opinion conclusion of the public opinion relationship map of the target project, For the reference of public opinion monitoring personnel of the target project. The public opinion conclusion includes positive public opinion information, negative public opinion information, event evaluation information and channel evaluation information of public opinion, wherein the positive public opinion information refers to the positive impact of public opinion, for example, for a mobile phone, the battery life of the mobile phone is long The negative public opinion information refers to the negative impact of public opinion. For example, for a mobile phone, the battery life of the mobile phone is short and the mobile phone is too hot when charging, etc. The event evaluation information refers to the impact of an event in the public opinion. Forecast evaluation and estimation, for example, the impact of a product evaluation report on the product, channel evaluation information refers to the impact of the channel to which the corpus source belongs to the target, for example, different websites have different audiences, scales, and influences, and events need to be evaluated Estimates of the influence of the channel to which the target belongs, for example, Weibo, WeChat circle of friends and forums have different influences on the target.

When combining the elements in the public opinion relationship map of the target item in a preset order to describe the public opinion of the target item in text form, the graphic database can be designed according to the relationship information between entities stored in the graphic database Information characteristics of time, distinguish the interrelationships between node sets, nodes and relationships in the graph, and then describe the relationship between nodes in words, so as to describe the public opinion of the target project in text form and give the public opinion of the project The monitoring staff will give textual prompts. For example, if in the public opinion relationship graph of the target project, there is a subordination relationship between nodes A and B, the public opinion of the target project can be described as "node A is subordinate to node B" when describing the public opinion of the target project in text form. Further, if the information that node A influences node B is obtained, the relevant information that node A influences node B can be further screened out from the acquired corpus, and the information that node A influences node B can be formed according to the trained regular expression or language model Abstract, provided in text form to the project public opinion monitoring personnel for reference, for example, the impact of the processor of the mobile phone on the mobile phone, or the impact of the battery of the mobile phone on the mobile phone, etc. Among them, a language model, such as an N-gram language model or a neural network language model.

Further, in one embodiment, by constructing the map data of the target project, a news corpus related to the target project is built. Before visualization, the public opinion data of the target project needs to be time-sorted, and the target projects are listed in chronological order Top-ranked news data to further filter out effective data and improve data processing efficiency. For example, by constructing the map data of a mobile phone, a news corpus related to the mobile phone is built. Before visualization, the public opinion data of the mobile phone needs to be time-sorted, and a list of the most recent rankings of the mobile phone is listed. News data to know the most concerned performance of this mobile phone recently, such as taking pictures or battery power.

In addition, in-depth analysis of the relevant fields of the target project can also be carried out. For example, in the product competition relationship, it is necessary to obtain public opinion data related to the target project from the attributes of competing products, and classify and deduplicate the data, and present it to users.

Further, to obtain the public opinion of the target project, realize the public opinion monitoring of the target project, and further deal with it according to the public opinion of the target project, realize the maintenance of the target project, for example, realize the public relations of the product image of the enterprise to maintain the image and image of the enterprise Benefit. For example, if the target project is a mobile phone, obtain the positive and negative information of the public opinion of the mobile phone, and obtain the event evaluation information and channel evaluation information in the public opinion of the mobile phone for corresponding public relations measures.

It should be noted that the project public opinion monitoring methods described in the above embodiments can recombine the technical features contained in different embodiments according to needs to obtain a combined implementation plan, but all of them are within the scope of protection required by this application within.

Please refer to FIG. 7, which is a schematic block diagram of a project public opinion monitoring device provided by an embodiment of the present application. Corresponding to the above project public opinion monitoring method, an embodiment of the present application further provides a project public opinion monitoring device. As shown in FIG. 7 , the project public opinion monitoring device includes a unit for executing the above-mentioned project public opinion monitoring method, and the device can be configured in a computer device such as a terminal. Specifically, please refer to FIG. 7 , the project public opinion monitoring device 700 includes a first acquisition unit 701 , a second acquisition unit 702 , a crawling unit 703 , an identification unit 704 , a construction unit 705 and a display unit 706 .

Wherein, the first acquiring unit 701 is configured to acquire the identification information of the target item in a preset manner;

The second acquiring unit 702 is configured to acquire a list of data source websites of the target item through a network search according to the identification information;

A crawling unit 703, configured to crawl the corpus of the target item from the data source websites included in the data source website list according to the identification information;

A recognition unit 704, configured to parse the corpus through natural language processing to identify subject names and public opinion features contained in the corpus;

A construction unit 705, configured to import the subject name and the public opinion features into the graph database to construct the public opinion relationship graph of the target project;

The display unit 706 is configured to display the public opinion relationship graph.

Please refer to FIG. 8 . FIG. 8 is another schematic block diagram of a project public opinion monitoring device provided by an embodiment of the present application. As shown in Figure 8, in this embodiment, the project public opinion monitoring device 700 also includes:

The update unit 707 is configured to update the list of data source websites by means of crawling.

Please continue to refer to FIG. 8, as shown in FIG. 8, the updating unit 707 includes:

The obtaining subunit 7071 is used to obtain the initial data source website list of the target project;

A classification subunit 7072, configured to classify the initial data source website list according to preset conditions to obtain different types of data source website lists;

Encapsulation subunit 7073, configured to encapsulate the different types of data source website lists into corresponding Docker containers;

The crawling subunit 7074 is used to start the Docker container so that the Docker container obtains a new data source website by crawling;

The update subunit 7075 is configured to add the new data source website according to type to the corresponding classified data source website list to update the data source website of the target project.

In one embodiment, the public opinion relationship graph of the target item includes the target item name, sub-item names of the target item, and public opinion features corresponding to the sub-item names.

Please refer to FIG. 8, as shown in FIG. 8, in this embodiment, the identification unit 704 includes:

The segmentation subunit 7041 is used to segment the corpus according to the sentence separator to obtain the sentence data set;

Build a subunit 7042, configured to build a named entity model according to the corpus;

The identification subunit 7043 is used to identify the subject names contained in the sentence data set through the named entity model;

The retrieval subunit 7044 is configured to perform part-of-speech analysis on the corpus and retrieve target relationships to obtain public opinion features of the target item.

Please refer to FIG. 8, as shown in FIG. 8, in this embodiment, the project public opinion monitoring device 700 also includes:

The description unit 708 is configured to combine elements in the public opinion relationship graph in a preset order to describe the public opinion of the target item in text.

In one embodiment, the description unit 708 is configured to combine the elements in the public opinion relationship graph in a preset order to describe the positive public opinion information, negative public opinion information, event evaluation information and channels of the target project in text form Assessment information.

It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above project public opinion monitoring device and each unit can refer to the corresponding descriptions in the foregoing method embodiments. Let me repeat.

At the same time, the division and connection of each unit in the above-mentioned project public opinion monitoring device are only for illustration. In other embodiments, the project public opinion monitoring device can be divided into different units according to needs, and each unit in the project public opinion monitoring device can also be The units adopt different connection sequences and methods to complete all or part of the functions of the public opinion monitoring device of the above project.

The above project public opinion monitoring device can be realized in the form of a computer program, and the computer program can run on the computer equipment as shown in FIG. 9 .

Please refer to FIG. 9. FIG. 9 is a schematic block diagram of a computer device provided by an embodiment of the present application. The computer device 900 may be a computer device such as a desktop computer or a server, or may be a component or component in other devices.

Referring to FIG. 9 , the computer device 900 includes a processor 902 connected through a system bus 901 , a memory and a network interface 905 , wherein the memory may include a non-volatile storage medium 903 and an internal memory 904 .

The non-volatile storage medium 903 can store an operating system 9031 and a computer program 9032 . When the computer program 9032 is executed, it can cause the processor 902 to execute the above-mentioned project public opinion monitoring method.

The processor 902 is used to provide calculation and control capabilities to support the operation of the entire computer device 900 .

The internal memory 904 provides an environment for the running of the computer program 9032 in the non-volatile storage medium 903. When the computer program 9032 is executed by the processor 902, the processor 902 can execute a method for monitoring project public opinion mentioned above.

The network interface 905 is used for network communication with other devices. Those skilled in the art can understand that the structure shown in FIG. 9 is only a block diagram of a partial structure related to the solution of this application, and does not constitute a limitation on the computer device 900 on which the solution of this application is applied. The specific computer device 900 may include more or fewer components than shown, or combine certain components, or have a different arrangement of components. For example, in some embodiments, the computer device may only include a memory and a processor. In such an embodiment, the structures and functions of the memory and the processor are consistent with those of the embodiment shown in FIG. 9 , and will not be repeated here.

Wherein, the processor 902 is configured to run a computer program 9032 stored in the memory to implement the following steps: obtain the identification information of the target item in a preset manner; obtain the target item through a network search according to the identification information A list of data source websites; crawl the corpus of the target project from the data source websites contained in the data source website list according to the identification information; analyze the corpus by natural language processing to identify the corpus contained in the corpus subject name and public opinion features; importing the subject name and the public opinion features into a graph database to construct a public opinion relationship map of the target project; displaying the public opinion relationship map.

In one embodiment, after the processor 902 implements the step of obtaining the list of data source websites of the target item by searching the Internet according to the identification information, it further implements the following steps:

The list of data source websites is updated by crawling.

In one embodiment, when the processor 902 implements the step of updating the list of data source websites by crawling, it specifically implements the following steps:

obtaining an initial data source website list of the target project;

Classifying the initial data source website list according to preset conditions to obtain different types of data source website lists;

Encapsulating the different types of data source website lists into corresponding Docker containers;

Start the Docker container so that the Docker container obtains the new data source website by crawling;

The new data source website is added to the corresponding classified data source website list according to type to update the data source website of the target project.

In one embodiment, when the processor 902 implements the public opinion relationship map of the target item, the public opinion relationship map specifically includes the following content: target item name, sub-item name of the target item, and sub-item name correspondence characteristics of public opinion.

In one embodiment, when the processor 902 implements the step of parsing the corpus through natural language processing to identify subject names and public opinion features contained in the corpus, the following steps are specifically implemented:

The corpus is divided according to the sentence separator to obtain a sentence data set;

Constructing a named entity model according to the corpus;

Identifying subject names contained in the sentence data set through the named entity model;

Perform part-of-speech analysis and target relationship retrieval on the corpus to obtain public opinion features of the target item.

In one embodiment, the processor 902 further implements the following steps after realizing the step of displaying the public opinion relationship graph:

Combining the elements in the public opinion relationship graph in a preset order to describe the public opinion of the target item in text form.

In one embodiment, when the processor 902 implements the step of combining elements in the public opinion relationship graph in a preset order to describe the public opinion of the target item in text form, the following steps are specifically implemented:

Combining the elements in the public opinion relationship graph in a preset order to describe the positive public opinion information, negative public opinion information, event evaluation information and channel evaluation information of the target project in text form.

It should be understood that in the embodiment of the present application, the processor 902 may be a central processing unit (Central Processing Unit, CPU), and the processor 902 may also be other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), dedicated integrated Circuit (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Wherein, the general-purpose processor may be a microprocessor, or the processor may be any conventional processor or the like.

Those skilled in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program, and the computer program can be stored in a computer-readable storage medium. The computer program is executed by at least one processor in the computer system to implement the process steps of the above method embodiments.

Therefore, the present application also provides a computer-readable storage medium. The computer-readable storage medium may be a non-volatile computer-readable storage medium, and the computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the processor performs the following steps:

A computer program product, when running on a computer, causes the computer to execute the steps of the project public opinion monitoring method described in the above embodiments.

The computer-readable storage medium may be an internal storage unit of the aforementioned device, such as a hard disk or memory of the device. The computer-readable storage medium can also be an external storage device of the device, such as a plug-in hard disk equipped on the device, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card , Flash Card (Flash Card) and so on. Further, the computer-readable storage medium may also include both an internal storage unit of the device and an external storage device.

Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described equipment, devices and units can refer to the corresponding process in the foregoing method embodiments, and details are not repeated here.

The computer-readable storage medium may be various computer-readable storage media capable of storing program codes, such as a U disk, a mobile hard disk, a read-only memory (Read-Only Memory, ROM), a magnetic disk, or an optical disk.

Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the relationship between hardware and software Interchangeability. In the above description, the composition and steps of each example have been generally described according to their functions. Whether these functions are executed by hardware or software depends on the specific application and design constraints of the technical solution. Skilled artisans may use different methods to implement the described functions for each specific application, but such implementation should not be regarded as exceeding the scope of the present application.

In the several embodiments provided in this application, it should be understood that the disclosed devices and methods may be implemented in other ways. For example, the device embodiments described above are illustrative only. For example, the division of each unit is only a logical function division, and there may be another division method in actual implementation. For example, several units or components may be combined or integrated into another system, or some features may be omitted, or not implemented.

The steps in the methods of the embodiments of the present application can be adjusted, combined and deleted according to actual needs. Units in the device in the embodiment of the present application may be combined, divided and deleted according to actual needs. In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, each unit may exist separately physically, or two or more units may be integrated into one unit.

If the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of this application is essentially or part of the contribution to the prior art, or all or part of the technical solution can be embodied in the form of software products, and the computer software products are stored in a storage medium In the above, several instructions are included to make an electronic device (which may be a personal computer, a terminal, or a network device, etc.) execute all or part of the steps of the methods described in the various embodiments of the present application.

The above is only a specific embodiment of the application, but the scope of protection of the invention is not limited thereto. Any person familiar with the technical field can easily think of various equivalents within the technical scope disclosed in this application. The modifications or replacements should be covered by the scope of protection of this application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims (10)

1. A project public opinion monitoring method, is characterized in that, described method comprises: Obtain the identification information of the target item through a preset method; Obtaining a list of data source websites of the target project through a network search according to the identification information; crawling the corpus of the target item from the data source websites included in the data source website list according to the identification information; Parsing the corpus by natural language processing to identify subject names and public opinion features contained in the corpus; Importing the subject name and the public opinion features into the graph database to construct the public opinion relationship graph of the target project; Display the public opinion relationship graph. 2. according to the described project public opinion monitoring method of claim 1, it is characterized in that, after the step of obtaining the data source website list of described target project by the mode of network search according to described identification information, also comprise: The list of data source websites is updated by crawling. 3. according to the described project public opinion monitoring method of claim 2, it is characterized in that, the described step of updating described data source website list by the mode of crawling comprises: obtaining an initial data source website list of the target project; Classifying the initial data source website list according to preset conditions to obtain different types of data source website lists; Encapsulating the different types of data source website lists into corresponding Docker containers; Start the Docker container so that the Docker container obtains the new data source website by crawling; The new data source website is added to the corresponding classified data source website list according to type to update the data source website of the target project. 4. The project public opinion monitoring method according to claim 1, wherein the public opinion relationship graph of the target project includes the name of the target project, the names of sub-projects of the target project, and public opinion features corresponding to the sub-project names. 5. according to the described project public opinion monitoring method of claim 1, it is characterized in that, the described corpus by natural language processing is analyzed to identify the main body title that described corpus contains and the step of public opinion feature comprises: The corpus is divided according to the sentence separator to obtain a sentence data set; Constructing a named entity model according to the corpus; Identifying subject names contained in the sentence data set through the named entity model; Perform part-of-speech analysis and target relationship retrieval on the corpus to obtain public opinion features of the target item. 6. The project public opinion monitoring method according to claim 1, characterized in that, after the step of displaying the public opinion relationship map, further comprising: Combining the elements in the public opinion relationship graph in a preset order to describe the public opinion of the target item in text form. 7. The project public opinion monitoring method according to claim 6, wherein the public opinion of the target project includes positive public opinion information, negative public opinion information, event evaluation information and channel evaluation information. 8. A project public opinion monitoring device, characterized in that it comprises: The first acquisition unit is configured to acquire the identification information of the target item in a preset manner; A second obtaining unit, configured to obtain a list of data source websites of the target project by searching the Internet according to the identification information; A crawling unit, configured to crawl the corpus of the target item from the data source websites included in the data source website list according to the identification information; A recognition unit, configured to parse the corpus through natural language processing to identify subject names and public opinion features contained in the corpus; A construction unit for importing the subject name and the public opinion features into the graph database to construct the public opinion relationship graph of the target project; A display unit, configured to display the public opinion relationship graph. 9. A computer device, characterized in that the computer device includes a memory and a processor connected to the memory; the memory is used to store a computer program; the processor is used to run the computer program stored in the memory , to execute the steps of the project public opinion monitoring method as described in any one of claims 1-7. 10. A computer-readable storage medium, characterized in that, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor executes any one of claims 1-7. The steps of the project public opinion monitoring method described in the item.