HK1175551B - Method and device for searching - Google Patents

Description

Searching method and device

Technical Field

The present application relates to the field of computer search technologies, and in particular, to a search method and apparatus.

Background

With the continuous development of computer network technology and wireless communication network technology, search services are more and more commonly applied, except that professional search websites are searched for the whole network, most websites also have a function of searching data contents in the website, especially for websites with rich data contents such as shopping websites, pages which users want to check are found out according to keywords input by the users, which is particularly important for the development of website services.

In a conventional search manner, a search server searches a large number of data objects (including but not limited to data contents in a website database such as a page, a data packet, etc.) in a website database according to keywords, finds data objects matching the keywords as search results, and presents the search results to a user.

In the scheme of obtaining search results in the search mode and displaying the search results in sequence, the search results are only sorted according to the literal relevance of the keywords, and some useful information can be arranged at the back position. For example, the keyword is "brand a mobile phone," the search server searches two search results, the first search result is a page for simply introducing the "brand a mobile phone," the second search result is a publication page containing images and texts of the "brand a mobile phone" and the "brand B mobile phone," only in terms of literal relevancy, the relevancy of the first search result is higher than that of the second search result, but in a specific website such as a shopping website, the second search result is richer than the content in the first search result, and the search requirement of the user can be met better. The ranking results obtained by ranking according to the literal relevancy of the keywords are not well matched with the search requirements of the user.

Therefore, on the basis of ranking according to the literal relevance of the keywords, a ranking scheme is further provided for mining the log information of the websites and indirectly representing the relevance according to the click rate of the user within a period of time, so that the search results arranged in the front are search results with high literal relevance of the keywords on one hand and search results with high relevance of the keywords on the other hand, namely search results which are likely to be interested by the user, in the search results of the search results finally displayed to the user, and the user can quickly find out useful information according to the ranked search results.

The two ways of sequencing the search results are performed according to the literal correlation and the usage correlation of the keywords, but in the business such as online shopping, when the sequencing is performed only according to the requirements of the two ways, the conditions on which the sequencing depends are limited, and the sequencing results can not meet the information query requirements of the user, so that after the search results obtained according to the sequencing way are pushed to the user, the user can only click and check the search results for many times until useful information is found. Because the sequencing of the search results output to the user at present has the problem of inaccurate sequencing results, the time overhead of searching information by the user is increased for the user, and the searching efficiency is reduced; on the other hand, for the web server, each time a user clicks one of the search results, the web server needs to respond to the clicking operation of the user, and if the user cannot quickly find out useful information from the search results arranged in the past, the web server needs to allocate a large amount of system resources for the operation of the user for clicking the search results multiple times, which causes waste of system resources.

Disclosure of Invention

The purpose of the embodiment of the application is as follows: the searching method and the searching equipment are provided for solving the problems that in the prior art, the searching efficiency is low, a large amount of system resources are allocated for the operation of clicking the searching result for multiple times, and the system resources are wasted.

A search method, comprising:

upon receiving a search request including a keyword, searching M search results from the database server using the keyword as a search condition, an

Determining N search results of which the transaction parameters corresponding to the received keywords meet set conditions according to the corresponding relation among the keywords, the search results and the transaction parameters of the search results;

and arranging N search results in front, arranging other search results except the search results which are the same as the N search results in the M search results, and returning the arranged search results, wherein M and N are positive integers.

A search apparatus, comprising:

the first search module is used for searching M search results from the database server by taking the keyword as a search condition when receiving a search request containing the keyword;

the second search module is used for determining N search results of which the transaction parameters corresponding to the received keywords meet the set conditions according to the corresponding relation among the keywords, the search results and the transaction parameters of the search results;

the sorting module is used for arranging the N search results in front of the search module and arranging other search results except the search results which are the same as the N search results in the M search results, wherein M and N are positive integers;

and the result returning module is used for returning the arranged search results.

The beneficial effect of this application is as follows:

in the scheme of the embodiment of the application, on one hand, the received keywords are taken as search conditions, and M search results with high correlation degree with the keywords are searched from the database server; on the other hand, according to the corresponding relation among the pre-learned keywords, the search results and the transaction parameters of the search results, N search results with higher transaction parameters are determined, the search results related to transaction conversion effects are arranged in front, the search results related to the relevance are arranged behind, and the search results arranged in front are the search results with high probability of generating purchasing behaviors, so that the probability that the search results arranged in front meet the query requirements of the user is higher, the search duration of the user can be effectively reduced, the search efficiency is improved, meanwhile, the user can complete the search process according to the search results arranged in front, the search results do not need to be clicked again, the system resources distributed by the network server for the click query operation of the user can be reduced, and the waste of the system resources is reduced; and the proposal can effectively improve the transaction conversion rate after the user searches, the user can select the commodity to be purchased only by trying few keywords, and the resources distributed by the server for the user searching process can be reduced because the keywords used by the user are few, so the limited resources of the server can support the searching access of more users.

Drawings

FIG. 1 is a schematic diagram illustrating steps of a search method according to an embodiment of the present disclosure;

fig. 2(a) and fig. 2(b) are schematic structural diagrams of a search apparatus in a second embodiment of the present application;

fig. 3 is a schematic diagram of a system architecture of a search device application according to a second embodiment of the present application.

Detailed Description

The embodiment of the application provides a new search scheme, and the search results are ranked by querying the contents of two aspects: on one hand, according to a conventional search mode, taking the received keywords as search conditions, and searching M search results with high correlation degrees with the keywords from a database server; on the other hand, N search results are determined according to the corresponding relation among the pre-learned keywords, the search results and the transaction parameters of the search results, and when the received keywords are taken as the search conditions of the N search results, the corresponding transaction parameters are higher, which means that the probability of purchasing behavior generated according to the N search results is higher. After the M search results and the N search results are found, the N search results related to the deal conversion effect are arranged in front as important search results, and other search results except the N search results are arranged in the M search results related to the relevancy. The method aims at providing the search results with high purchasing behavior probability for users after being inquired by most users within the preset time length when the arranged search results are displayed to the users, so that the probability that the arranged search results meet the inquiry requirements of the users is higher, the search time length of the users can be effectively reduced, the search efficiency is improved, and meanwhile, the users do not need to click the search results displayed by the inquiry for many times, the system resources distributed by the network server for the click inquiry operation of the users can be reduced, and the waste of the system resources is reduced; and the proposal can effectively improve the transaction conversion rate after the user searches, the user can select the commodity to be purchased only by trying few keywords, and the resources distributed by the server for the user searching process can be reduced because the keywords used by the user are few, so the limited resources of the server can support the searching access of more users.

The search results referred to in the embodiments of the present application refer to: data objects in the website database, such as pages, data packets, etc. in the website.

The bargaining parameters related in the embodiments of the application refer to: when a user obtains a certain search result under the condition that the specific keyword is taken as a search condition, the user queries the search result and finally generates the probability of purchasing behavior, namely the transaction conversion effect after the user queries the specific search result under the specific keyword.

The embodiments of the present application will be described in detail below with reference to the drawings attached hereto.

Example one

As shown in fig. 1, a schematic diagram of steps of a search method in an embodiment of the present application is shown, where the method includes the following steps:

step 101: a search request is received that includes a keyword (query).

In the scheme of the step, when a user wants to perform a search operation, a search request is initiated and keywords serving as search conditions are carried in the search request.

Step 102: and searching M search results from the database server by taking the keyword as a search condition, wherein M is a positive integer.

And reading the keywords in the search requests, establishing links with the database server, and searching for M search requests with higher correlation degrees with the keywords from the database server, wherein the M search results are higher correlation degrees with the keywords and therefore may be the search results which the user wishes to view.

The searching process in this step may be a process of searching according to the degree of literal relevance to the keyword, or may be a process of searching according to the degree of literal relevance to the keyword, and further according to mining of log information of the website, searching out the search result in which the user is interested. It should be noted that the present step is not limited to the above two search manners, and all other manners capable of searching out a search result according to a keyword can be applied to the present step.

Step 103: and determining N search results of which the deal parameters corresponding to the received keywords meet set conditions according to the corresponding relation among the keywords, the search results and the deal parameters of the search results, wherein N is a positive integer.

When a search request is received, on one hand, M search results related to the keyword are searched from the database server in the mode of step 102, and on the other hand, N search results, in which the deal parameters corresponding to the received keyword meet the set conditions, are determined according to the corresponding relationship among the keyword, the search results and the deal parameters of the search results, which is obtained after local learning. In this step, the N search results may be determined from a database server according to the correspondence, or the N search results may be determined from locally stored search results according to the correspondence.

The N search results obtained in this step are search results with a high deal conversion rate, that is, within the set time length before, the probability that the N search results generate purchasing behavior after being queried by the user is high. Because the N search results can well meet the query and purchase requirements of the user within the set time length, the probability that the N search results can meet the query and purchase requirements of the user sending the search request at this time is high.

It should be noted that, in the first embodiment of the present application, the execution order of step 102 and step 103 may not be fixed, step 103 may be executed first, step 102 is executed later, or step 102 and step 103 may be executed simultaneously.

Step 104: and arranging the N search results in front, and arranging the other search results except the search results which are the same as the N search results in the M search results.

Since the N search results obtained in step 103 are queried under the condition of the deal parameter, compared with the M search results searched under the condition of the correlation with the keyword in step 102, the N search results can reflect the correlation with the keyword to some extent and can also represent the usability of the search results themselves, so the N search results can reflect the query requirements of the user better, and the N search results are ranked in front of the M search results in this step.

For example: the user logs in a shopping website and inputs 'brand A mobile phone' as a keyword to initiate a search request in a search dialog box of the website. After receiving the search request, 3 search results, namely search result 1, search result 2 and search result 3, are searched out in a database server of the website under the condition of 'brand A mobile phone'. Meanwhile, according to the search results obtained after other users search by taking the brand-A mobile phone as the keyword within the previous 1 week and the behaviors of inquiring and purchasing the obtained search results by the users under the condition that the brand-A mobile phone is the keyword, 2 search results with higher transaction parameters are finally determined, namely a search result 2 and a search result 4 respectively. At this time, the search results 2 and 4 are search results of multi-user query and purchase, so the search results 2 and 4 are available and valid search results and should be ranked in the front; although the relevance of the search result 1, the search result 2 and the search result 3 to the keywords is high, the search results themselves may have usability and effectiveness problems, such as unrealistic situations that the quality of the linked commodities of the search result 1 is poor, the price is too high or too low, and the like, and therefore, the search result 1 and the search result 3 should be ranked behind.

Because the search results stored in the database server are continuously changed, the search results stored in the current database server are not always identical to the search results used for learning to obtain the corresponding relation within the preset time length; meanwhile, the N search results determined according to the correspondence are determined according to the size of the deal parameter, and the M search results searched in step 102 are determined according to the correlation with the keyword, and therefore, it is likely that the M search results in step 102 and the N search results in step 103 are not completely identical.

However, since the N search results may also store search results having a higher degree of correlation with the keyword, the N search results and the M search results may also overlap, and in order to avoid repeated display of the search results, in this step, when ranking, the N search results that do not overlap with each other are arranged in front, and then the search results that are the same as the N search results in the M search results are removed, and the remaining search results are arranged behind. Under a special condition, if the N search results and the M search results are completely the same, the N search results or the M search results may be directly used as the search results.

Step 105: and returning the arranged search results to the user equipment initiating the search request.

After the search results are ranked in step 104, the ranked search results may be returned to the user equipment in a list manner, for example, a summary or a page of link addresses of the ranked search results may be pushed to the user equipment.

According to the scheme of the first embodiment of the application, the search results with high transaction conversion rate are arranged in front, that is, the search results with high availability and effectiveness are arranged in the front of the search result display page, and because the probability that the search results which are more front are browsed and inquired by the user is higher, the user who initiates the search request can browse and inquire the search results with high transaction conversion rate first, so that the user can search out useful information quickly, the search efficiency of the user is improved, and system resources which are distributed by the network server for the user and used for inquiring are reduced. In addition, because the user firstly inquires the search result with high transaction conversion rate, and the possibility that the user generates purchasing behavior aiming at the inquired search result is high, the transaction conversion rate aiming at the search process is also improved, the user can select the commodity to be purchased only by trying few keywords, and because the keywords used by the user are few, the resources distributed by the server to the search process of the user can be reduced, so that the search access of more users can be supported under the condition that the server resources are the same.

The following is a detailed description of each step of the examples of the present application.

In the first embodiment of the present application, the correspondence among the keywords, the search results, and the deal parameters of the search results, which are used in step 103, is determined after a large number of users query the search results within a preset time period and learn about purchasing behavior of the search results, and the determination manner of the correspondence is specifically described below:

the first step is as follows: and collecting search requests containing keywords sent by each user within a set time length.

The set time in this step can be set according to the need, if the data with more abundant data volume for determining the corresponding relation is needed, the set time can be set for a longer time; if the latest data for determining the corresponding relationship is required to be obtained, the set time length may be set to a shorter time, and the acquired data may be periodically updated.

The second step is that: and respectively searching from the database server by taking each keyword as a search condition, and determining a search result group which corresponds to each keyword and contains at least one search result.

Suppose that the keywords received in the set time length are brand a and cell phone. And (3) searching by taking the brand A as a searching condition to obtain a searching result group 1, wherein the searching results comprise: "a cell phone", "a display", and "a game machine"; searching by taking the 'mobile phone' as a search condition to obtain a search result group 2, wherein the search results comprise: "A cell phone", "B cell phone" and "cell phone accessory".

The third step: and determining a click index of the search result and a purchase index generated after clicking the search result when the keyword corresponding to the search result group is taken as a search condition for each search result.

Taking the search result "mobile phone a" as an example, the search result group where the "mobile phone a" is located is the search result group 1 and the search result group 2, and when the user obtains the search result "mobile phone a" with the keyword "brand a" corresponding to the search result group 1 as a search condition, determining the click index and purchase index of the "mobile phone a" under the condition. When the user obtains the search result 'a mobile phone' with the keyword 'mobile phone' corresponding to the search result group 2 as the search condition, the click index and purchase index of the 'a mobile phone' under the condition are determined. And by analogy, the click index and the purchase index of each search result in the second step under the specific keyword are obtained.

The click index in the present embodiment is information indicating the number of times a certain search condition is queried under a specific keyword, that is, the number of times a search result is queried obtained when a keyword corresponding to a search result group in which the search result is located is used as the search condition. The queried number information includes the number of users who query the search result or the number of times the search result is queried. For example: when 50 users click and inquire the obtained search result 'brand A mobile phone' by taking 'brand A' as a search condition, the search is clicked and inquired for 200 times, and then the click index of 'brand A mobile phone' under the keyword 'brand A' can be preset to be 50 users or 200 click and inquire numbers.

The purchase index in this embodiment is information indicating the number of times that a certain search condition is queried under a specific keyword to generate a purchase behavior, that is, when the keyword corresponding to the search result group in which the search result is located is used as the search condition, the obtained search result is queried by the user, and then the user purchases a corresponding product according to the query result. The information of the times of generating the purchasing behavior comprises the number of users generating the purchasing behavior or the times of generating the purchasing behavior according to the search result. For example: when 50 users click and inquire the obtained search result 'brand A mobile phone' by taking 'brand A' as a search condition, the click and inquiry are carried out for 200 times, 8 users click and inquire and then purchase the brand A mobile phone, the 8 users purchase the brand A mobile phone for 12 times, and the purchase index of the 'brand A mobile phone' under the keyword 'brand A' is the number of purchased users 8 or the accumulated purchase number 12.

According to the above manner of generating the click index and the purchase index, the click index and the purchase index of each search result under a specific search term in the second step can be obtained, the click index and the purchase index of each search result under one search term are recorded as a field, and the field at least comprises: < keyword >, < search result >, < click index of search result under the keyword >, < purchase index of search result under the keyword >. Combining the click indexes and the fields of the purchase indexes of all the search results collected in the second step under the respective keywords together to form a list structure shown in table 1, wherein the list structure shown in table 1 takes the number of times of the search results being queried as the click index of the search results, and the number of times of the purchase behavior generated according to the queried search results as the purchase index of the search results.

Keyword	Search results	Click index	Purchase indicator
				Brand A	A mobile phone	200	12
Brand A	A display	150	6
				Brand A	A game machine	300	9
Mobile phone	A mobile phone	200	8
				Mobile phone	B mobile phone	500	50
Mobile phone	Mobile phone accessory	250	10

TABLE 1

The fourth step: and generating a transaction parameter according to the click index and the purchase index, wherein the transaction parameter is used as the transaction parameter corresponding to the search result when the keyword corresponding to the search result group is used as the search condition.

After the list structure shown in table 1 is obtained, the click index and the purchase index in each field in table 1 may be calculated to obtain the deal parameters of the search result under the specific keyword. The manner of calculating the transaction parameters includes, but is not limited to, the following three manners and variations of the three manners:

the first way to calculate the transaction parameters:

generating a transaction parameter according to formula (1):

wherein Z represents a transaction parameter of the search result; x represents a purchase index of the search result; y represents the click index of the search result.

The second way to calculate the transaction parameters:

generating a transaction parameter according to formula (2):

wherein, X^/Representing the variance of the purchase indicators for all search results within the search result set in which the search result is located.

The third way to calculate the transaction parameters:

generating a transaction parameter according to formula (3):

wherein, Y^/Representing the variance of the click index for all search results within the search result set in which the search result is located.

The purchase index in the above formulas (1) to (3) may be the number of users who have made a purchase, or the number of times a purchase has been made; the click index may be the number of users who have queried the search result or the number of times the search result has been queried. When calculating the deal parameters of each search result, as long as the meaning of the purchase index and the click index used in the calculation of each search result is the same, the calculated deal parameters can reflect the deal conversion rate of the search result under the specific keyword. For example, for each search result, the used purchase index is the number of times of generating purchase behavior and the click index is the number of times of querying the search result, or the used purchase index is the number of users generating purchase behavior and the click index is the number of times of querying the search result.

Assuming that the purchase index in this embodiment is the number of times of generating a purchase behavior, and the click index is the number of times of querying a search result, the list structure shown in table 1 may be transformed into the list structure shown in table 2, and at this time, a corresponding relationship among a keyword, a search result, and a deal parameter of the search result is obtained.

Keyword	Search results	Parameters of transaction
			Brand A	A mobile phone	0.06
Brand A	A display	0.04
			Brand A	A game machine	0.03
Mobile phone	A mobile phone	0.04
			Mobile phone	B mobile phone	0.1
Mobile phone	Mobile phone accessory	0.04

TABLE 2

In step 103, when a search request including the keyword "brand a" is received, a deal parameter of different search results under the keyword can be obtained through the lookup table 2, and then the search result of which the deal parameter reaches the set threshold value is used as N search results of which the deal parameter satisfies the set conditions, and subsequent operations are performed.

In the manner of generating the correspondence among the keywords, the search results, and the deal parameters of the search results in this embodiment, various source data acquisition means may be employed to obtain data required in the first to third steps, for example, the source data is obtained after log information is analyzed, and then the list structure shown in table 2 is obtained.

In step 104 of the first embodiment, the specific implementation manner of step 104 is different according to different situations of the ordering of the M search results and the N search results themselves.

The M search results obtained in the first embodiment may be arranged in the order of highest relevance, or may be in the case of no arrangement but carrying the relevance score of each search result. The relevancy score represents the relevancy of the search result and the keyword, and the relevancy score corresponding to the search result with higher relevancy is larger.

Similarly, the N search results may be arranged in an order from big to small in the transaction parameter, or in a case where there is no arrangement order but the transaction amount score of the search result is carried. The deal amount score represents the deal conversion rate of the search result, and the larger the deal parameter, the larger the deal amount score corresponding to the search result.

The M search results and the N search results are different in order, and the implementation manner of step 104 in the first embodiment is also not completely the same, which is described below:

the first case: the M search results are arranged according to the sequence of the relevance from high to low, and the N search results are arranged according to the sequence of the traffic parameters from large to small.

For the first case, the specific implementation manner of step 104 is:

since the N search results and the M search results are all sorted, the N search results that have been sorted in the order of highest transaction parameter to lowest can be directly sorted in front, and the other search results except the N search results among the M search results can be sorted in the order of highest relevance to the keyword to lowest.

It should be noted that the sorting manner in the first case is also applicable to all other cases, that is, when M search results are not sorted, the sorting operation may be performed according to the relevancy score of each search result; when the N search results are not sorted, the search results may also be sorted according to the volume score of each search result to obtain the sorted M search results and N search results, and at this time, all other situations are converted into the first situation.

The second case: the M search results are not provided with relevancy scores and are sorted in the order of highest relevancy to lowest relevancy, and the N search results are provided with volume of transaction scores and are not sorted.

For the second case, the specific implementation manner of step 104 is:

after the N search results are ranked according to the sequence of the volume of interest scores from large to small, the second situation is converted into the first situation, the N search results which are ranked according to the sequence of the volume of interest scores from high to low can be directly ranked in front, and the other search results except the N search results in the M search results are ranked in the sequence of the relevance of the search results to the keywords from high to low.

The third situation: the M search results are provided with the relevancy scores but are not sorted, and the N search results are not provided with the volume of transaction scores and are arranged according to the sequence of transaction parameters from large to small.

For the third case, the specific implementation manner of step 104 is:

firstly, taking M search results as search results in a queue to be processed.

The search results in the queue to be processed are the search results which are finally output to the user, and the queue to be processed can be regarded as an intermediate queue for integrating the M search results and the N search results.

Assume that M search results in this step are: "a cell phone", "a computer", and "a game machine", at this time, the contents in the queue to be processed are: [ mobile phone A (relevance score 1), computer A (relevance score 2), game machine A (relevance score 3) ].

And then, sequentially reading each search result in the N search results according to the corresponding transaction parameters from big to small.

Suppose that the N search results in this step are arranged in the order of big to small in the transaction parameters: "a cell phone", "a display", and "a game machine". The following operations are executed in a loop for each read search result until all the N search results have been read:

and reading the 1 st search result 'A mobile phone' in the N search results, and judging whether the search result 'A mobile phone' exists in the queue to be processed. Because the search result of the 'mobile phone a' exists in the queue to be processed, the search result of the 'mobile phone a' is a search result with a high transaction conversion rate and a search result with a high keyword relevance degree, and such a search result is a result which most possibly meets the search requirement of the user, therefore, the sum of the relevance degree score 1 of the 'mobile phone a' and the weight value 1 of the search result of the 'mobile phone a' in the N search results is used as the relevance degree score of the 'mobile phone a' in the queue to be processed.

At this time, the contents in the queue to be processed are updated as follows: [ mobile phone (relevance score 1+ weight 1), computer (relevance score 2), game machine (relevance score 3) ].

And reading the 2 nd search result display A in the N search results, writing the display A into the queue to be processed because the display A does not exist in the queue to be processed, and taking the weighted value 2 of the display A in the N search results as the relevancy score of the display A.

At this time, the contents in the queue to be processed are updated as follows: [ mobile phone A (relevance score 1+ weight value 1), computer A (relevance score 2), game machine A (relevance score 3), and display A (weight value 2) ].

Reading the 3 rd search result 'A game' in the N search results for the third time, and taking the sum of the relevance score 3 of the 'A game' and the weight value 3 of the search result 'A game' in the N search results as the relevance score of the 'A mobile phone' in the queue to be processed because the search result 'A game' exists in the queue to be processed.

At this time, the contents in the queue to be processed are updated as follows: [ mobile phone A (relevance score 1+ weight value 1), computer A (relevance score 2), game machine A (relevance score 3+ weight value 3), and display A (weight value 2) ].

The weight values in the N search results may be pre-assigned, wherein the larger the transaction parameter is, the larger the assigned weight value is, that is, for the N search results that have been arranged in the order of the transaction parameter from large to small, the weight value assigned to the search result arranged in the front is greater than the weight value assigned to the search result arranged in the back.

And finally, after the N search results are completely read, arranging the search results in the queue to be processed in the order of the relevance scores from high to low.

In a third scenario, the relevancy scores of the search results in the pending queue may also be obtained by weighting the relevancy scores and the weight values.

A fourth scenario: the M search results are not ranked with relevance scores and the N search results are not ranked with volume scores.

For the fourth scenario, the specific implementation manner of step 104 is:

firstly, taking M search results as search results in a queue to be processed.

Still assume that the M search results in this step are: "a cell phone", "a computer", and "a game machine", at this time, the contents in the queue to be processed are: [ mobile phone A (relevance score 1), computer A (relevance score 2), game machine A (relevance score 3) ].

The content of the N search results is [ a mobile phone (volume of transaction score 1), a display (volume of transaction score 2), and a game machine (volume of transaction score 3) ], and at this time, the arrangement order of the N search results is not according to the arrangement order of the volume of transaction scores from large to small.

Then, each of the N search results is read (each of the N search results may be read in any order), and the following is performed for each read search result in a loop until all of the N search results have been read:

and reading the 1 st search result 'A mobile phone' in the N search results, and judging whether the search result 'A mobile phone' exists in the queue to be processed. Because the search result of the 'A mobile phone' exists in the queue to be processed, the sum of the relevance score 1 of the 'A mobile phone' and the traffic score 1 of the 'A mobile phone' is used as the relevance score of the 'A mobile phone' in the queue to be processed.

At this time, the contents in the queue to be processed are updated as follows: [ mobile phone A (relevance score 1+ volume score 1), computer A (relevance score 2), game machine A (relevance score 3) ].

And reading the 2 nd search result display A in the N search results, writing the display A into the queue to be processed because the display A does not exist in the queue to be processed, and taking the trading volume score 2 of the display A as the relevancy score of the display A.

At this time, the contents in the queue to be processed are updated as follows: [ A Mobile phone (relevance score 1+ volume score 1), A computer (relevance score 2), A game machine (relevance score 3), A display (volume score 2) ].

And reading the 3 rd search result 'A game machine' in the N search results for the third time, wherein the search result 'A game machine' exists in the queue to be processed, and therefore the sum of the relevance score 3 of the 'A game machine' and the traffic score 3 of the 'A game machine' is used as the relevance score of the 'A mobile phone' in the queue to be processed.

At this time, the contents in the queue to be processed are updated as follows: [ A Mobile phone (relevance score 1+ volume score 1), A computer (relevance score 2), A game machine (relevance score 3+ volume score 3), A display (volume score 2) ].

And finally, after the N search results are completely read, arranging the search results in the queue to be processed in the order of the relevance scores from high to low.

It should be noted that, in the fourth scenario, the relevancy score of the search results in the queue to be processed may also be obtained by weighting the relevancy score and the volume score.

In the first to fourth cases, if the number of search results in the queue to be processed is large, only the optimal search results may be displayed. Before the search results are displayed to the user, a plurality of optimal search results in the queue to be processed can be sorted and then output to the display device for displaying; the optimal search results in the queue to be processed can also be not sorted, but the search results and the relevancy scores are sent to the display device, and the display device sorts and displays the received search results according to the received relevancy scores.

Example two

As shown in fig. 2(a), a schematic structural diagram of a search apparatus in the second embodiment of the present application includes a first search module 11, a second search module 12, a sorting module 13, and a result returning module 14, where: the first search module 11 is configured to, when receiving a search request including a keyword, search M search results from the database server by using the keyword as a search condition; the second search module 12 is configured to determine, according to the correspondence among the keyword, the search result, and the deal parameters of the search result, N search results in which the deal parameters corresponding to the received keyword satisfy the set conditions; the sorting module 13 is configured to arrange the N search results in front of the search result, and arrange other search results, which are the same as the N search results, of the M search results after the search results are divided, where M and N are positive integers; the result returning module 14 is used for returning the arranged search results.

The first search module 11 may be a search engine for searching.

The system further comprises a relation establishing module 15, which is used for collecting the received search requests containing the keywords within the set duration, respectively searching from the database server by taking each keyword as a search condition, determining a search result group containing at least one search result corresponding to each keyword, determining a click index of the search result and a purchase index generated after clicking the search result when the keyword corresponding to the search result group is taken as the search condition for each search result, and generating a transaction parameter according to the click index and the purchase index, wherein the transaction parameter is taken as a transaction parameter corresponding to the search result when the keyword corresponding to the search result group is taken as the search condition.

Table 2 generated in embodiment one may be stored in the relationship establishing module 15.

Aiming at different situations of sequencing of M search results and N search results obtained by the first search module and the second search module, the sequencing modes of the sequencing modules are different:

for the first situation and the second situation in the first embodiment, the sorting module 13 is specifically configured to, if the M search results are arranged from high to low in correlation with the keyword, and the deal parameters corresponding to the N search results are arranged from large to small, arrange the N search results in front, and arrange the other search results, except the N search results, in the M search results in the order from high to low in correlation with the keyword.

As shown in fig. 2(b), for the third case in the first embodiment, the sorting module 13 includes a score determining sub-module 21, a reading sub-module 22, a judging sub-module 23, and an operating sub-module 24, where: the score determining submodule 21 is configured to determine a relevance score of each of the M search results; the reading submodule 22 is configured to sequentially read each search result of the N search results according to the corresponding transaction parameters in a descending order until all the N search results are read; the judgment submodule 23 is configured to use the M search results as search results in a queue to be processed, and judge whether there is an ith read search result in the queue to be processed, and if there is no ith read search result, use a weight value assigned to the ith search result as a relevance score of the search result, and write the relevance score into the queue to be processed; if yes, taking the sum of the weight value distributed to the ith search result and the relevance grade of the ith search result as a new relevance grade of the ith search result, wherein i is a positive integer with the value of 1-N; the operation submodule 24 is configured to, after the judgment submodule judges all the read N search results, arrange the search results in the queue to be processed in an order from high to low according to the relevancy scores.

For the fourth case in the first embodiment, the score determining sub-module 21 is configured to determine a relevancy score of each of the M search results and a deal score of each of the N search results; the reading sub-module 22 is configured to read each search result of the N search results until all of the N search results are read; the judgment submodule 23 is configured to use the M search results as search results in the queue to be processed, and judge whether there is an ith read search result in the queue to be processed, and if there is no ith read search result, use a traffic score of the ith search result as a relevancy score of the search result, and write the relevancy score into the queue to be processed; if yes, taking the sum of the volume score of the ith search result and the relevancy score of the ith search result as a new relevancy score of the ith search result, wherein i is a positive integer with the value of 1-N; the operation submodule 24 is configured to, after the judgment submodule judges all the read N search results, arrange the search results in the queue to be processed in an order from high to low according to the relevancy scores.

The device for sorting search results in the second embodiment has a functional module for executing each step in the first embodiment, and may be software, hardware, or a combination of the software and the hardware, which is not described herein again.

The search device in the second embodiment may be applied to the system architecture shown in fig. 3, where the system architecture shown in fig. 3 includes a user device, a search device, and a database server, where: after the user equipment sends a search request containing the key words to the search equipment, the search equipment searches from the database server by taking the key words as search conditions, and after the scheme in the embodiment of the application is carried out on the searched search results, the search results are returned to the user equipment, so that the search requirements of the user equipment are met. The search device in this embodiment is not limited to be used in other system architectures.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

While the preferred embodiments of the present application have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all alterations and modifications as fall within the scope of the application.

It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.

Claims (8)

1. A method of searching, comprising:

upon receiving a search request including a keyword, searching M search results from the database server using the keyword as a search condition, an

Determining N search results of which the transaction parameters of the search results corresponding to the received keyword meet set conditions according to the corresponding relation among the keyword, the search results and the transaction parameters of the search results; the corresponding relation among the keywords, the search results and the transaction parameters of the search results is pre-established in the following mode: collecting search requests containing keywords received within a set time length, respectively searching from a database server by taking each keyword as a search condition, and determining a search result group containing at least one search result corresponding to each keyword; aiming at each search result, determining a click index of the search result and a purchase index generated after clicking the search result when a keyword corresponding to the search result group is taken as a search condition; generating a transaction parameter according to the click index and the purchase index, wherein the transaction parameter is used as the transaction parameter corresponding to the search result when the keyword corresponding to the search result group is used as the search condition;

and arranging N search results in front, arranging other search results except the search result which is the same as the N search results in the M search results in back, and returning the arranged search results, wherein M and N are positive integers.

2. The method of claim 1, wherein the click metrics for the search results are: when the keyword corresponding to the search result group in which the search result is located is taken as a search condition, the number of users inquiring the search result, or the number of times the search result is inquired;

the purchase indexes are as follows: and the number of users who generate purchasing behaviors according to the search result or the times of generating purchasing behaviors according to the search result.

3. The method of claim 2, wherein the deal parameters are generated according to the following formula:

wherein Z represents a transaction parameter; x represents the purchase index of the search result corresponding to the transaction parameter; y represents a click index of the deal parameter corresponding to the search result; or

Generating a transaction parameter according to the following formula:

wherein Z represents a transaction parameter; x represents the purchase index of the search result corresponding to the transaction parameter; y represents a click index of the deal parameter corresponding to the search result; x^/Representing the variance of the purchase indexes of all the search results in the search result group of the corresponding search result of the transaction parameters; or

Generating a transaction parameter according to the following formula:

wherein Z represents a transaction parameter; x represents the purchase index of the search result corresponding to the transaction parameter; y represents a click index of the deal parameter corresponding to the search result; y is^/And expressing the variance of the click indexes of all the search results in the search result group of the corresponding search result of the deal forming parameter.

4. The method according to claim 1, wherein if the M search results are arranged from high to low in the correlation with the keyword and the transaction parameters corresponding to the N search results are arranged from large to small, the N search results are arranged in front, and the other search results except the search result same as the N search results in the M search results are arranged in back in the order from high to low in the correlation with the keyword.

5. The method of claim 1, wherein the method further comprises:

determining a relevance score for each of the M search results;

the arranging N search results in front and arranging other search results, except the search result the same as the N search results, in the M search results in the back specifically includes:

taking the M search results as search results in a queue to be processed, sequentially reading each search result in the N search results according to the corresponding transaction parameters from big to small, and executing the following operations aiming at each read search result:

judging whether an ith read search result exists in a queue to be processed, wherein i is a positive integer with the value of 1-N;

if not, taking the weight value distributed to the ith search result as the relevancy score of the search result, and writing the relevancy score into a queue to be processed; if yes, taking the sum of the weight value distributed to the ith search result and the relevance grade of the ith search result as a new relevance grade of the ith search result;

after the read N search results are all subjected to the operation, the search results in the queue to be processed are arranged in the order of the relevancy scores from high to low.

6. The method of claim 5, wherein the larger the corresponding deal parameter of the N search results, the larger the weight value assigned.

7. The method of claim 1, wherein the method further comprises:

determining a relevancy score of each of the M search results and a volume score of each of the N search results;

the arranging N search results in front and arranging other search results, except the search result the same as the N search results, in the M search results in the back specifically includes:

taking the M search results as search results in a queue to be processed, reading each search result in the N search results, and executing the following operations aiming at each read search result:

judging whether an ith read search result exists in a queue to be processed, wherein i is a positive integer with the value of 1-N;

if not, taking the traffic score of the ith search result as the relevancy score of the search result, and writing the relevancy score into a queue to be processed; if yes, taking the sum of the volume score of the ith search result and the relevancy score of the ith search result as a new relevancy score of the ith search result;

after the read N search results are all subjected to the operation, the search results in the queue to be processed are arranged in the order of the relevancy scores from high to low.

8. A search apparatus, comprising:

the relation establishing module is used for establishing the corresponding relation among the keywords, the search results and the transaction parameters of the search results in advance in the following modes: collecting search requests containing keywords received within a set time length, respectively searching from a database server by taking each keyword as a search condition, determining a search result group containing at least one search result corresponding to each keyword, determining a click index of the search result and a purchase index generated after clicking the search result when the keyword corresponding to the search result group is taken as the search condition aiming at each search result, and generating a transaction parameter according to the click index and the purchase index to serve as the transaction parameter corresponding to the search result when the keyword corresponding to the search result group is taken as the search condition;

the first search module is used for searching M search results from the database server by taking the keyword as a search condition when receiving a search request containing the keyword;

the second search module is used for determining N search results of which the transaction parameters of the search results corresponding to the received keywords meet set conditions according to the corresponding relation among the keywords, the search results and the transaction parameters of the search results;

the sorting module is used for arranging N search results in front of the search results and arranging other search results except the search results which are the same as the N search results in the M search results behind the search results, wherein M and N are positive integers;

and the result returning module is used for returning the arranged search results.