CN112000871A - Method, device and equipment for determining search result list and storage medium - Google Patents
Method, device and equipment for determining search result list and storage medium Download PDFInfo
- Publication number
- CN112000871A CN112000871A CN202010850973.XA CN202010850973A CN112000871A CN 112000871 A CN112000871 A CN 112000871A CN 202010850973 A CN202010850973 A CN 202010850973A CN 112000871 A CN112000871 A CN 112000871A
- Authority
- CN
- China
- Prior art keywords
- search
- information
- search result
- sample
- recommendation effect
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 230000000694 effects Effects 0.000 claims abstract description 130
- 239000013598 vector Substances 0.000 claims description 38
- 238000012163 sequencing technique Methods 0.000 claims description 13
- 230000015654 memory Effects 0.000 claims description 9
- 230000006870 function Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 5
- 238000009825 accumulation Methods 0.000 description 4
- 230000000007 visual effect Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 230000001960 triggered effect Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000001186 cumulative effect Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9532—Query formulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9538—Presentation of query results
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application discloses a method, a device, equipment and a storage medium for determining a search result list, and belongs to the technical field of computers. The method comprises the following steps: obtaining a plurality of search results corresponding to the target search keyword; determining different sorting information corresponding to the plurality of search results arranged in different orders; for each piece of ranking information, determining a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword and a recommendation effect scoring model; and determining a search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results. The recommendation effect of the search result list can be improved through the method and the device.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a method, an apparatus, a device, and a storage medium for determining a search result list.
Background
When people use computer equipment daily, a search function is inevitably used, when a user searches, a terminal can display a page of a search result list, the page of the search result list comprises a plurality of search results, and each search result in the page of the search result list is displayed according to a certain sequence, and the specific processing is as follows:
after the server receives the keyword input by the user, the server can acquire the search result containing the keyword according to the keyword, sort the search results based on the historical click times of the search results to generate sorting information, and then determine a search result list corresponding to the keyword based on the sorting information and the search results.
In the process of implementing the present application, the inventor finds that the prior art has at least the following problems:
the historical click times are greatly influenced by the visual attractiveness of each search result, for example, a search result containing videos and pictures is more visually attractive than a search result only containing characters, however, the visual attraction is strong and does not mean that the search result containing the video and the picture better meets the requirements of the user, when the user clicks, may click on the search result only according to the visual appeal of the search result, which also makes the historical click times of the search result containing video and pictures high, therefore, the historical number of clicks does not represent how well the search result meets the user's needs, which reflects only the degree of visual appeal to the user, and, therefore, in the search result list obtained from the historical number of clicks, the arrangement order of the search results is not arranged according to the degree of meeting the requirements of the user, which results in poor recommendation effect of the search result list.
Disclosure of Invention
The embodiment of the application provides a method, a device, equipment and a storage medium for determining a search result list, and can solve the problem that the recommendation effect of the search result list is not good. The technical scheme is as follows:
in one aspect, a method of determining a search result list is provided, the method comprising:
obtaining a plurality of search results corresponding to the target search keyword;
determining different sorting information corresponding to the plurality of search results arranged in different orders;
for each piece of ranking information, determining a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword and a recommendation effect scoring model;
and determining a search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results.
Optionally, before obtaining the plurality of search results corresponding to the target search keyword, the method further includes:
receiving a search request which is sent by a terminal and carries a target search keyword;
after determining the search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results, the method further includes:
and sending the search result list to the terminal.
Optionally, the search request further carries a user identifier, and the determining of the recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword, and the recommendation effect scoring model includes:
and determining a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword, the user information corresponding to the user identification and a recommendation effect scoring model.
Optionally, the obtaining of the plurality of search results corresponding to the target search keyword includes:
when a preset updating period is reached, acquiring a plurality of search results corresponding to the target search keyword;
after determining the search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results, the method further includes:
and updating the locally stored search result list corresponding to the target search keyword based on the determined search result list corresponding to the target search keyword.
Optionally, after updating the locally stored search result list corresponding to the target search keyword based on the determined search result list corresponding to the target search keyword, the method further includes:
and when a search request which is sent by a terminal and carries a target search keyword is received, sending a search result list corresponding to the locally stored target search keyword to the terminal.
Optionally, before obtaining the plurality of search results corresponding to the target search keyword, the method further includes:
obtaining a sample search keyword and a plurality of sample search results corresponding to the sample search keyword;
determining sample ordering information for the plurality of sample search results;
determining a reference recommendation effect score corresponding to the sample sorting information based on the historical click times of the sample search keywords corresponding to each sample search result and the sample sorting information;
and training the recommendation effect scoring model based on the sample sorting information, the format information of each sample search result, the sample search keywords and the reference recommendation effect score.
Optionally, for each piece of ranking information, determining a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword, and the recommendation effect scoring model includes:
generating a feature vector corresponding to each search result based on the format information of each search result and the target search keyword;
combining the feature vectors corresponding to all the search results into a total feature vector according to the sequence of the search results in the sorting information;
and inputting the total feature vector into the recommendation effect scoring model to obtain a recommendation effect score corresponding to the sequencing information.
In another aspect, an apparatus for determining a search result list is provided, the apparatus comprising:
the acquisition module is used for acquiring a plurality of search results corresponding to the target search keyword;
the determining module is used for determining different sequencing information corresponding to the plurality of search results arranged according to different sequences;
the scoring module is used for determining a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword and a recommendation effect scoring model for each ranking information;
and the list determining module is used for determining a search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results.
Optionally, the apparatus further includes a sending module, where the sending module is configured to:
receiving a search request which is sent by a terminal and carries a target search keyword;
after determining the search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results, the method further includes:
and sending the search result list to the terminal.
Optionally, the search request further carries a user identifier, and the scoring module is configured to:
and determining a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword, the user information corresponding to the user identification and a recommendation effect scoring model.
Optionally, the obtaining module is configured to:
when a preset updating period is reached, acquiring a plurality of search results corresponding to the target search keyword;
after determining the search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results, the method further includes:
and updating the locally stored search result list corresponding to the target search keyword based on the determined search result list corresponding to the target search keyword.
Optionally, the apparatus further includes a training module, where the training module is configured to:
obtaining a sample search keyword and a plurality of sample search results corresponding to the sample search keyword;
determining sample ordering information for the plurality of sample search results;
determining a reference recommendation effect score corresponding to the sample sorting information based on the historical click times of the sample search keywords corresponding to each sample search result and the sample sorting information;
and training the recommendation effect scoring model based on the sample sorting information, the format information of each sample search result, the sample search keywords and the reference recommendation effect score.
Optionally, the scoring module is configured to:
generating a feature vector corresponding to each search result based on the format information of each search result and the target search keyword;
combining the feature vectors corresponding to all the search results into a total feature vector according to the sequence of the search results in the sorting information;
and inputting the total feature vector into the recommendation effect scoring model to obtain a recommendation effect score corresponding to the sequencing information.
In yet another aspect, a computer device is provided, the computer device comprising a processor and a memory, the memory having stored therein instructions, execution of which by the processor causes the computer device to implement the method of determining a search result list.
In yet another aspect, a computer-readable storage medium is provided, which stores instructions that, when executed by a computer device, cause the computer device to implement the method of determining a search result list.
The technical scheme provided by the embodiment of the application has the following beneficial effects:
according to the method and the device, the format information of each search result corresponding to the target search keyword, the target search keyword and different corresponding sorting information are input into the recommendation effect scoring model, recommendation effect scores of a plurality of search results arranged according to the different sorting information are obtained, the search result list corresponding to the target search keyword is determined according to the sorting information with the highest recommendation effect score and the plurality of search results, and the search result list obtained through the scheme is the search result list which is sorted according to the different format information of each search result and the relevance of the keyword, so that the recommendation effect of the search result list is improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a schematic diagram of an implementation environment provided by an embodiment of the present application;
FIG. 2 is a flowchart of a method for determining a target search result list according to an embodiment of the present application;
FIG. 3 is a schematic structural diagram of an apparatus for determining a target search result list according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of a server according to an embodiment of the present application.
Detailed Description
To make the objects, technical solutions and advantages of the present application more clear, embodiments of the present application will be described in further detail below with reference to the accompanying drawings.
The embodiment of the application provides a method for determining a target search result list, which may be implemented by a server, where the server may be a background server of the application program or the web page, and the server may be an individual server or a server group, and if the server is an individual server, the server may be responsible for all processing that needs to be performed by the server in the following scheme, and if the server is a server group, different servers in the server group may be respectively responsible for different processing in the following scheme, and specific processing allocation conditions may be arbitrarily set by a technical person according to actual needs, and are not described herein again.
As shown in fig. 1, when a user uses a terminal, the user often searches, for example, by browsing a search web page through a browser and performing a search, or by performing a search in an application. When searching, a user can input a keyword to be searched in an input box, then click a search control behind the input box, and then the terminal can send a search request to the server, wherein the search request contains the keyword, and the server obtains a plurality of corresponding search results and sequencing information of the search results according to the keyword, so that the search results and the sequencing sequence are sent to the terminal.
Fig. 2 is a flowchart of a method for determining a target search result list according to an embodiment of the present application. Referring to fig. 2, the process includes:
In implementation, the server may be triggered to obtain the contents of the plurality of search results corresponding to the target search keyword in a plurality of ways, which may specifically include the following ways:
in the first mode, when a preset updating period is reached, the contents of a plurality of search results corresponding to a target search keyword are obtained.
In practice, the developer may set the update period for the server, so that the server has a preset update period, for example, the preset update period is one day. And then, when the server detects that the time length after the last updating is finished reaches a preset updating period, the server acquires the contents of a plurality of search results containing the target search keywords.
Optionally, before performing the above processing, the server may train the recommendation effect scoring model.
First, a sample search keyword and contents of a plurality of sample search results corresponding to the sample search keyword are obtained.
In implementation, the server obtains the sample search keyword from the stored search keywords, and obtains the contents of a plurality of sample search results corresponding to the sample search keyword in the historical search records, wherein the contents of the plurality of sample search results are arranged according to the sample ordering information.
Second, sample ordering information for a plurality of sample search results is determined.
In an implementation, sample ordering information for the plurality of sample search results is determined according to an order of arrangement of the plurality of sample search results.
And thirdly, determining a reference recommendation effect score corresponding to the sample sorting information based on the historical click times of the sample search keywords corresponding to each sample search result and the sample sorting information.
In implementation, the server obtains the content of each sample search result obtained by searching based on the sample search keyword, and further obtains the historical click times of the sample search keyword corresponding to each sample search result, and further calculates the reference recommendation effect score corresponding to the sample ranking information based on the historical click times, the sample ranking information and the normalized breaking and loss accumulation gain function.
For example, the formula of the normalized break-up cumulative gain function is: NDCG (NDCG)k＝DCGk/IDCGk
Wherein the content of the first and second substances,
IDCGkrel for re-ordered scores based on historical click timesiRepresenting the historical number of clicks for the ith search result.
After the historical click times of the sample search results are obtained, the server inputs the historical click times and the sample sorting information into the DCG of the normalized breaking accumulation gain functionkIn the formula, a first fraction DCG of the first search result list is calculatedkThen, the contents of the sample search result are reordered based on the historical click times to obtain rearranged sample ordering information, and then a second score IDCG is calculated based on the rearranged sample ordering information and the historical click times of the sample search resultkThen using the first fraction DCGkDivide by a second fraction IDCGkAnd obtaining the reference recommendation effect score corresponding to the sample sorting information.
And then training a recommendation effect scoring model based on the sample sorting information, the format information of each sample search result, the sample search keywords and the reference recommendation effect score.
The format information comprises picture information, video information and character information.
In implementation, the server may input the sample ranking information, the format information of each sample search result, the sample search keywords, the user information corresponding to the user identifier, and the reference recommendation effect score into the recommendation effect scoring model. The recommendation effect scoring model firstly extracts the features of the format information of the sample search results, the sample search keywords and the user information corresponding to the user identification to obtain a 256-dimensional vector corresponding to the sample search results, then combines a plurality of 256-dimensional vectors corresponding to the contents of the sample search results into a target vector according to the sample sorting information, and then calculates the training recommendation effect score according to the target vector and the dynamic neural network. Then, based on the training recommendation effect score, the benchmark recommendation effect score, and the loss function
And adjusting the recommendation effect scoring model.
And then, repeating the processing until the training of the recommendation effect scoring model is finished.
In the second mode, the terminal can send a search request carrying the target search keyword to the server, so that the server obtains the contents of a plurality of search results corresponding to the target search keyword after receiving the search request carrying the target search keyword sent by the terminal.
In implementation, the terminal may display a page of the application program, an input box may be displayed in the page, and then the user may input the target search keyword in the input box and trigger the search control, and then the terminal may send a search request carrying the target search keyword to the server. The server receives the search request carrying the target search keyword, acquires the target search keyword in the search request, and acquires the content of a plurality of search results containing the target search keyword.
Optionally, before performing the above processing, the server may train the recommendation effect scoring model.
First, a sample search keyword and contents of a plurality of sample search results corresponding to the sample search keyword are obtained.
In implementation, the server obtains the sample search keyword from the stored search keywords, and obtains the contents of a plurality of sample search results corresponding to the sample search keyword in the historical search records, wherein the contents of the plurality of sample search results are arranged according to the sample ordering information.
Second, sample ordering information for a plurality of sample search results is determined.
In an implementation, sample ordering information for the plurality of sample search results is determined according to an order of arrangement of the plurality of sample search results.
And thirdly, determining a reference recommendation effect score corresponding to the sample sorting information based on the historical click times of the sample search keywords corresponding to each sample search result, the user information corresponding to the user identification and the sample sorting information.
In implementation, the server obtains the content of each sample search result obtained by searching based on the sample search keyword, and further obtains the historical click times of the sample search keyword corresponding to each sample search result, and further calculates the reference recommendation effect score corresponding to the sample ranking information based on the historical click times, the sample ranking information and the normalized breaking and loss accumulation gain function.
For example, the formula of the normalized break-up cumulative gain function is: NDCG (NDCG)k＝DCGk/IDCGk
Wherein the content of the first and second substances,
IDCGkrel for re-ordered scores based on historical click timesiRepresenting the historical number of clicks for the ith search result.
After the historical click times of the sample search results are obtained, the server inputs the historical click times and the sample sorting information into the DCG of the normalized breaking accumulation gain functionkIn the formula, a first fraction DCG of the content of a plurality of search results arranged according to the sample ordering information is calculatedkThen, the contents of the sample search result are reordered based on the historical click times to obtain rearranged sample ordering information, and then a second score IDCG is calculated based on the rearranged sample ordering information and the historical click times of the sample search resultkThen using the first fraction DCGkDivide by a second fraction IDCGkAnd obtaining the reference recommendation effect score corresponding to the sample sorting information.
And then training a recommendation effect scoring model based on the sample sorting information, the format information of each sample search result, the sample search keywords, the user information corresponding to the user identification and the reference recommendation effect score.
The format information comprises picture information, video information and character information.
In an implementation, the server may order the sample ordering information, each sampleAnd inputting format information of the search result, the sample search keyword, user information corresponding to the user identification and the reference recommendation effect score into the recommendation effect scoring model. The recommendation effect scoring model firstly extracts the format information of the sample search results, the sample search keywords and the user information corresponding to the user identification to obtain a 256-dimensional feature vector corresponding to the sample search results, then combines a plurality of 256-dimensional vectors corresponding to the content of the sample search results into a target vector according to the sample sorting information, and then calculates the training recommendation effect score according to the target vector and the dynamic neural network. Then, based on the training recommendation effect score, the benchmark recommendation effect score, and the loss function
And adjusting the recommendation effect scoring model.
And then, repeating the processing until the training of the recommendation effect scoring model is finished.
In implementation, after obtaining the contents of the plurality of search results corresponding to the target search keyword, the contents of the plurality of search results may be arranged in different orders, so as to obtain different ranking information.
And 203, for each piece of ranking information, determining a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword and a recommendation effect scoring model.
In implementation, after obtaining different ranking information, the server may sequentially input the content of the search results arranged according to the different ranking information into the recommendation effect scoring model, and the specific processing may be as follows:
first, a feature vector corresponding to the content of each search result is generated based on the format information of each search result and the target search keyword.
And the recommendation effect scoring model performs feature extraction on the format information and the target search keywords of each search result to obtain 256-dimensional feature vectors corresponding to each search result.
And secondly, combining the feature vectors corresponding to the contents of all the search results into a total feature vector according to the sequence of the search results in the sequencing information.
And then, inputting the total feature vector into a recommendation effect scoring model to obtain a recommendation effect score corresponding to the sequencing information.
Optionally, if step 201 is triggered by the terminal sending a plurality of search requests, the processing of step 203 may determine a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword, the user information corresponding to the user identifier, and the recommendation effect scoring model. The user identifier is carried in the search request, and the specific processing may be as follows:
first, user information corresponding to a user identifier is obtained based on the user identifier.
And secondly, generating a feature vector corresponding to each search result based on the format information of each search result, the target search keyword and the user information corresponding to the user identification.
And the recommendation effect scoring model performs feature extraction on the format information of each search result, the target search keyword and the user information corresponding to the user identification to obtain 256-dimensional feature vectors corresponding to each search result.
And thirdly, combining the feature vectors corresponding to the contents of all the search results into a total feature vector according to the sequence of the search results in the sequencing information.
And then, inputting the total feature vector into a recommendation effect scoring model to obtain a recommendation effect score corresponding to the sequencing information.
It should be noted that step 203 is repeated until all the contents of the search results arranged according to the different ranking information are calculated to obtain the recommendation effect score.
And step 204, determining a search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the contents of the plurality of search results.
In implementation, after the processing of step 203, the server may obtain the recommendation effect scores corresponding to the plurality of ranking information, and further, the server may obtain the ranking information with the highest recommendation effect score and the contents of the plurality of search results, and generate a search result list corresponding to the target search keyword according to the ranking information with the highest recommendation effect score and the contents of the plurality of search results.
Optionally, if step 201 is triggered by the terminal sending a plurality of search requests, the search result list may be sent to the terminal after the processing of step 204 is completed.
Optionally, if the step 201 is performed based on a preset update period, after the processing of the step 204 is completed, the search result list corresponding to the target search keyword stored locally may be updated based on the search result list corresponding to the determined target search keyword.
Correspondingly, after updating, when the server receives a search request which is sent by the terminal and carries a target search keyword, a search result list corresponding to the target search keyword which is locally stored can be sent to the terminal.
According to the method and the device, the format information of each search result corresponding to the target search keyword, the target search keyword and different corresponding sorting information are input into the recommendation effect scoring model, recommendation effect scores of a plurality of search results arranged according to the different sorting information are obtained, the search result list corresponding to the target search keyword is determined according to the sorting information with the highest recommendation effect score and the plurality of search results, and the search result list obtained through the scheme is the search result list which is sorted according to the different format information of each search result and the relevance of the keyword, so that the recommendation effect of the search result list is improved.
All the above optional technical solutions may be combined arbitrarily to form the optional embodiments of the present disclosure, and are not described herein again.
Fig. 3 is a schematic structural diagram of an apparatus for determining a target search result list according to an embodiment of the present application. The apparatus may be the server described above. Referring to fig. 3, the apparatus includes:
an obtaining module 310, configured to obtain a plurality of search results corresponding to a target search keyword;
a determining module 320, configured to determine that the plurality of search results are arranged in different orders and corresponding to different ordering information;
the scoring module 330 is configured to determine, for each piece of ranking information, a recommendation effect score corresponding to the ranking information based on the ranking information, format information of each search result, the target search keyword, and a recommendation effect scoring model;
the list determining module 340 is configured to determine a search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results.
Optionally, the apparatus further includes a sending module, where the sending module is configured to:
receiving a search request which is sent by a terminal and carries a target search keyword;
after determining the search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results, the method further includes:
and sending the search result list to the terminal.
Optionally, the search request further carries a user identifier, and the scoring module 330 is configured to:
and determining a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword, the user information corresponding to the user identification and a recommendation effect scoring model.
Optionally, the obtaining module 310 is configured to:
when a preset updating period is reached, acquiring a plurality of search results corresponding to the target search keyword;
after determining the search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results, the method further includes:
and updating the locally stored search result list corresponding to the target search keyword based on the determined search result list corresponding to the target search keyword.
Optionally, the apparatus further includes a training module, where the training module is configured to:
obtaining a sample search keyword and a plurality of sample search results corresponding to the sample search keyword;
determining sample ordering information for the plurality of sample search results;
determining a reference recommendation effect score corresponding to the sample sorting information based on the historical click times of the sample search keywords corresponding to each sample search result and the sample sorting information;
and training the recommendation effect scoring model based on the sample sorting information, the format information of each sample search result, the sample search keywords and the reference recommendation effect score.
Optionally, the scoring module 330 is configured to:
generating a feature vector corresponding to each search result based on the format information of each search result and the target search keyword;
combining the feature vectors corresponding to all the search results into a total feature vector according to the sequence of the search results in the sorting information;
and inputting the total feature vector into the recommendation effect scoring model to obtain a recommendation effect score corresponding to the sequencing information.
According to the method and the device, the format information of each search result corresponding to the target search keyword, the target search keyword and different corresponding sorting information are input into the recommendation effect scoring model, recommendation effect scores of a plurality of search results arranged according to the different sorting information are obtained, the search result list corresponding to the target search keyword is determined according to the sorting information with the highest recommendation effect score and the plurality of search results, and the search result list obtained through the scheme is the search result list which is sorted according to the different format information of each search result and the relevance of the keyword, so that the recommendation effect of the search result list is improved.
It should be noted that: in the apparatus for determining a search result list according to the above embodiment, when determining a search result list, only the division of the above functional modules is used as an example, and in practical applications, the above function distribution may be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules, so as to complete all or part of the above described functions. In addition, the embodiments of the method for determining the search result list provided by the above embodiments belong to the same concept, and specific implementation processes thereof are described in the embodiments of the method for determining the search result list, which are not described herein again.
Fig. 4 is a schematic structural diagram of a server according to an embodiment of the present application, where the server may be the above server, and the server 400 may generate a relatively large difference due to a difference in configuration or performance, and may include one or more processors (CPUs) 401 and one or more memories 402, where at least one instruction is stored in the memory 1002, and the at least one instruction is loaded and executed by the processor 401 to implement the methods provided by the above method embodiments. Of course, the server may also have components such as a wired or wireless network interface, a keyboard, and an input/output interface, so as to perform input/output, and the server may also include other components for implementing the functions of the device, which are not described herein again.
In an exemplary embodiment, a computer-readable storage medium, such as a memory, including instructions executable by a processor in a terminal to perform the method of determining a search result list in the above embodiments is also provided. For example, the computer-readable storage medium may be a Read-only Memory (ROM), a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
The above description is only exemplary of the present application and should not be taken as limiting, as any modification, equivalent replacement, or improvement made within the spirit and principle of the present application should be included in the protection scope of the present application.
Claims (15)
1. A method of determining a search result list, the method comprising:
obtaining a plurality of search results corresponding to the target search keyword;
determining different sorting information corresponding to the plurality of search results arranged in different orders;
for each piece of ranking information, determining a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword and a recommendation effect scoring model;
and determining a search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results.
2. The method of claim 1, wherein before obtaining the plurality of search results corresponding to the target search keyword, the method further comprises:
receiving a search request which is sent by a terminal and carries a target search keyword;
after determining the search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results, the method further includes:
and sending the search result list to the terminal.
3. The method according to claim 2, wherein the search request further carries a user identifier, and the determining of the recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword, and the recommendation effect scoring model comprises:
and determining a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword, the user information corresponding to the user identification and a recommendation effect scoring model.
4. The method of claim 1, wherein the obtaining of the plurality of search results corresponding to the target search keyword comprises:
when a preset updating period is reached, acquiring a plurality of search results corresponding to the target search keyword;
after determining the search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results, the method further includes:
and updating the locally stored search result list corresponding to the target search keyword based on the determined search result list corresponding to the target search keyword.
5. The method according to claim 4, wherein after updating the locally stored search result list corresponding to the target search keyword based on the determined search result list corresponding to the target search keyword, the method further comprises:
and when a search request which is sent by a terminal and carries a target search keyword is received, sending an updated search result list corresponding to the locally stored target search keyword to the terminal.
6. The method of claim 1, wherein before obtaining the plurality of search results corresponding to the target search keyword, the method further comprises:
obtaining a sample search keyword and a plurality of sample search results corresponding to the sample search keyword;
determining sample ordering information for the plurality of sample search results;
determining a reference recommendation effect score corresponding to the sample sorting information based on the historical click times of the sample search keywords corresponding to each sample search result and the sample sorting information;
and training the recommendation effect scoring model based on the sample sorting information, the format information of each sample search result, the sample search keywords and the reference recommendation effect score.
7. The method according to claim 1, wherein for each piece of ranking information, determining a recommendation effect score corresponding to the ranking information based on the ranking information, format information of each search result, the target search keyword, and a recommendation effect scoring model comprises:
generating a feature vector corresponding to each search result based on the format information of each search result and the target search keyword;
combining the feature vectors corresponding to all the search results into a total feature vector according to the sequence of the search results in the sorting information;
and inputting the total feature vector into the recommendation effect scoring model to obtain a recommendation effect score corresponding to the sequencing information.
8. An apparatus for determining a search result list, the apparatus comprising:
the acquisition module is used for acquiring a plurality of search results corresponding to the target search keyword;
the determining module is used for determining different sequencing information corresponding to the plurality of search results arranged according to different sequences;
the scoring module is used for determining a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword and a recommendation effect scoring model for each ranking information;
and the list determining module is used for determining a search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results.
9. The apparatus of claim 8, further comprising a sending module configured to:
receiving a search request which is sent by a terminal and carries a target search keyword;
after determining the search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results, the method further includes:
and sending the search result list to the terminal.
10. The apparatus of claim 9, wherein the search request further carries a user identifier, and the scoring module is configured to:
and determining a recommendation effect score corresponding to the ranking information based on the ranking information, the format information of each search result, the target search keyword, the user information corresponding to the user identification and a recommendation effect scoring model.
11. The apparatus of claim 8, wherein the obtaining module is configured to:
when a preset updating period is reached, acquiring a plurality of search results corresponding to the target search keyword;
after determining the search result list corresponding to the target search keyword based on the ranking information with the highest recommendation effect score and the plurality of search results, the method further includes:
and updating the locally stored search result list corresponding to the target search keyword based on the determined search result list corresponding to the target search keyword.
12. The apparatus of claim 8, further comprising a training module to:
obtaining a sample search keyword and a plurality of sample search results corresponding to the sample search keyword;
determining sample ordering information for the plurality of sample search results;
determining a reference recommendation effect score corresponding to the sample sorting information based on the historical click times of the sample search keywords corresponding to each sample search result and the sample sorting information;
and training the recommendation effect scoring model based on the sample sorting information, the format information of each sample search result, the sample search keywords and the reference recommendation effect score.
13. The apparatus of claim 8, wherein the scoring module is configured to:
generating a feature vector corresponding to each search result based on the format information of each search result and the target search keyword;
combining the feature vectors corresponding to all the search results into a total feature vector according to the sequence of the search results in the sorting information;
and inputting the total feature vector into the recommendation effect scoring model to obtain a recommendation effect score corresponding to the sequencing information.
14. A computer device, comprising a processor and a memory, wherein at least one instruction is stored in the memory, and is loaded and executed by the processor to perform operations performed by the method of determining a search result list according to any one of claims 1 to 7.
15. A computer-readable storage medium having stored therein at least one instruction which is loaded and executed by a processor to perform operations performed by the method of determining a search result list according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010850973.XA CN112000871A (en) | 2020-08-21 | 2020-08-21 | Method, device and equipment for determining search result list and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010850973.XA CN112000871A (en) | 2020-08-21 | 2020-08-21 | Method, device and equipment for determining search result list and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112000871A true CN112000871A (en) | 2020-11-27 |
Family
ID=73473975
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010850973.XA Withdrawn CN112000871A (en) | 2020-08-21 | 2020-08-21 | Method, device and equipment for determining search result list and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112000871A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112818262A (en) * | 2021-01-28 | 2021-05-18 | 上海博泰悦臻网络技术服务有限公司 | Map POI searching method, system, device and medium based on user data |
| CN113521758A (en) * | 2021-08-04 | 2021-10-22 | 北京字跳网络技术有限公司 | Information interaction method and device, electronic equipment and storage medium |
| CN113672700A (en) * | 2021-08-18 | 2021-11-19 | 北京达佳互联信息技术有限公司 | Content item searching method and device, electronic equipment and storage medium |
| WO2022262849A1 (en) * | 2021-06-17 | 2022-12-22 | 浙江口碑网络技术有限公司 | Search result output method and apparatus, computer device and readable storage medium |
| WO2023231288A1 (en) * | 2022-06-02 | 2023-12-07 | 北京百度网讯科技有限公司 | Sorting method and apparatus for search results, and electronic device and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160055252A1 (en) * | 2014-05-07 | 2016-02-25 | Yandex Europe Ag | Methods and systems for personalizing aggregated search results |
| US20170075897A1 (en) * | 2015-09-14 | 2017-03-16 | Yandex Europe Ag | System and method for ranking search results based on usefulness parameter |
| CN108345601A (en) * | 2017-01-23 | 2018-07-31 | 腾讯科技（深圳）有限公司 | Search result ordering method and device |
| CN109299344A (en) * | 2018-10-26 | 2019-02-01 | Oppo广东移动通信有限公司 | The generation method of order models, the sort method of search result, device and equipment |
| CN109582852A (en) * | 2018-12-05 | 2019-04-05 | 中国银行股份有限公司 | A kind of sort method and system of full-text search result |
| CN110321474A (en) * | 2019-05-21 | 2019-10-11 | 北京奇艺世纪科技有限公司 | Recommended method, device, terminal device and storage medium based on search term |
-
2020
- 2020-08-21 CN CN202010850973.XA patent/CN112000871A/en not_active Withdrawn
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160055252A1 (en) * | 2014-05-07 | 2016-02-25 | Yandex Europe Ag | Methods and systems for personalizing aggregated search results |
| US20170075897A1 (en) * | 2015-09-14 | 2017-03-16 | Yandex Europe Ag | System and method for ranking search results based on usefulness parameter |
| CN108345601A (en) * | 2017-01-23 | 2018-07-31 | 腾讯科技（深圳）有限公司 | Search result ordering method and device |
| CN109299344A (en) * | 2018-10-26 | 2019-02-01 | Oppo广东移动通信有限公司 | The generation method of order models, the sort method of search result, device and equipment |
| CN109582852A (en) * | 2018-12-05 | 2019-04-05 | 中国银行股份有限公司 | A kind of sort method and system of full-text search result |
| CN110321474A (en) * | 2019-05-21 | 2019-10-11 | 北京奇艺世纪科技有限公司 | Recommended method, device, terminal device and storage medium based on search term |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112818262A (en) * | 2021-01-28 | 2021-05-18 | 上海博泰悦臻网络技术服务有限公司 | Map POI searching method, system, device and medium based on user data |
| CN112818262B (en) * | 2021-01-28 | 2023-07-21 | 上海博泰悦臻网络技术服务有限公司 | Map POI searching method, system, equipment and medium based on user data |
| WO2022262849A1 (en) * | 2021-06-17 | 2022-12-22 | 浙江口碑网络技术有限公司 | Search result output method and apparatus, computer device and readable storage medium |
| CN113521758A (en) * | 2021-08-04 | 2021-10-22 | 北京字跳网络技术有限公司 | Information interaction method and device, electronic equipment and storage medium |
| CN113521758B (en) * | 2021-08-04 | 2023-10-24 | 北京字跳网络技术有限公司 | Information interaction method and device, electronic equipment and storage medium |
| CN113672700A (en) * | 2021-08-18 | 2021-11-19 | 北京达佳互联信息技术有限公司 | Content item searching method and device, electronic equipment and storage medium |
| CN113672700B (en) * | 2021-08-18 | 2023-10-13 | 北京达佳互联信息技术有限公司 | Content item searching method, device, electronic equipment and storage medium |
| WO2023231288A1 (en) * | 2022-06-02 | 2023-12-07 | 北京百度网讯科技有限公司 | Sorting method and apparatus for search results, and electronic device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112000871A (en) | Method, device and equipment for determining search result list and storage medium | |
| CN109086439A (en) | Information recommendation method and device | |
| CN109168047B (en) | Video recommendation method and device, server and storage medium | |
| WO2007130914A2 (en) | Dynamic search engine results employing user behavior | |
| CN112818224B (en) | Information recommendation method and device, electronic equipment and readable storage medium | |
| CN111105269B (en) | Advertisement delivery processing method, device, equipment and storage medium | |
| CN111737443B (en) | Answer text processing method and device and key text determining method | |
| CN112784141A (en) | Search result quality determination method and device, storage medium and computer equipment | |
| CN112860929A (en) | Picture searching method and device, electronic equipment and storage medium | |
| CN110264283B (en) | Popularization resource display method and device | |
| CN116186326A (en) | Video recommendation method, model training method, electronic device and storage medium | |
| CN111506596A (en) | Information retrieval method, information retrieval device, computer equipment and storage medium | |
| CN102760127B (en) | Method, device and the equipment of resource type are determined based on expanded text information | |
| CN115935049A (en) | Recommendation processing method and device based on artificial intelligence and electronic equipment | |
| CN112115354A (en) | Information processing method, information processing apparatus, server, and storage medium | |
| CN109299378B (en) | Search result display method and device, terminal and storage medium | |
| CN111597469A (en) | Display position determining method and device, electronic equipment and storage medium | |
| CN111797765B (en) | Image processing method, device, server and storage medium | |
| CN115878874A (en) | Multimodal retrieval method, device and storage medium | |
| CN110110202A (en) | A kind of information flow method for pushing and device | |
| CN114461822A (en) | Resource processing method, device, equipment and storage medium | |
| CN114491093A (en) | Multimedia resource recommendation and object representation network generation method and device | |
| CN113704623A (en) | Data recommendation method, device, equipment and storage medium | |
| CN113515701A (en) | Information recommendation method and device | |
| CN111127079A (en) | Method, device, computer equipment and storage medium for issuing commodity resources |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20201127 |
|
| WW01 | Invention patent application withdrawn after publication |