Target search method and device, electronic equipment and readable storage medium

By constructing and rewriting search terms, combining user portraits and historical search logs, the problem of excessively broad results of traditional search systems is solved, and more accurate and personalized search results are achieved, improving user experience.

CN119046523APending Publication Date: 2024-11-29BEIJING LONGZHI DIGITAL TECH CO LTD
View PDF 0 Cites 4 Cited by

Patent Information

Application Number
CN202411146901.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-20
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

传统搜索系统返回的搜索结果过于宽泛,难以满足用户的个性化搜索需求。

Method used

By obtaining the search terms entered by the user and the user's portrait information, constructing the first search terms based on the user's historical search log, and obtaining the rewritten search terms based on the scene information, searching to obtain the target search results.

Benefits of technology

It improves the accuracy and efficiency of search results, meets users' personalized search needs, and improves user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119046523A_ABST
    Figure CN119046523A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of artificial intelligence, and provides a target searching method and device, electronic equipment and a readable storage medium. The method comprises the steps of obtaining a search term input by a user; constructing a first search term by utilizing the search term and user portrait information, and obtaining a first search result corresponding to the first search term; constructing a second search term by utilizing the first search term, the first search result and a historical search log of the user; and based on the second search term and the scene information, obtaining a rewritten search term, and performing search by using the rewritten search term to obtain a target search result corresponding to the rewritten search term. According to the method and the device, personalized search requirements of the user can be met, search efficiency and search accuracy are improved, and meanwhile, search experience of the user is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a target search method, device, electronic device and readable storage medium. Background Art

[0002] In the context of the digital age, search technology has developed into a key bridge connecting information and user needs. In activities such as online shopping, obtaining the latest information or enjoying entertainment content, the search system has become an indispensable auxiliary tool for users. However, traditional search systems mainly search based on the original meaning of the search terms entered by users or based on simple expanded terms. This method is relatively fixed, and the search results returned are often too broad and difficult to accurately match the user's actual search intentions. Therefore, this search method based solely on search terms or expansions can no longer meet the personalized search needs of users. Summary of the invention

[0003] In view of this, the embodiments of the present application provide a target search method, device, electronic device and readable storage medium to solve the problem that the search results returned by the prior art are often too broad and cannot meet the user's personalized search needs.

[0004] In a first aspect of an embodiment of the present application, a target search method is provided, the method comprising:

[0005] Obtain the search term input by the user; construct a first search term using the search term and user portrait information, and obtain a first search result corresponding to the first search term; construct a second search term using the first search term, the first search result, and the user's historical search log; determine the scenario information corresponding to the second search term, obtain a rewritten search term based on the second search term and the scenario information, and use the rewritten search term to search to obtain a target search result corresponding to the rewritten search term.

[0006] According to a second aspect of the embodiments of the present application, a target search device is provided, the method comprising:

[0007] An acquisition module is configured to acquire a search term input by a user; a first construction module is configured to construct a first search term using the search term and user portrait information, and obtain a first search result corresponding to the first search term; a second construction module is configured to construct a second search term using the first search term, the first search result, and the user's historical search log; a search module is configured to determine scenario information corresponding to the second search term, obtain a rewritten search term based on the second search term and the scenario information, and use the rewritten search term to search to obtain a target search result corresponding to the rewritten search term.

[0008] According to a third aspect of an embodiment of the present application, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0009] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, which stores a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0010] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0011] Obtain the search term input by the user; construct a first search term using the search term and user portrait information, and obtain a first search result corresponding to the first search term; construct a second search term using the first search term, the first search result, and the user's historical search log; determine the scene information corresponding to the second search term, obtain a rewritten search term based on the second search term and the scene information, and use the rewritten search term to search and obtain a target search result corresponding to the rewritten search term. Construct a first search term set based on the search term input by the user and the user portrait information, and conduct a preliminary search based on it to obtain a first search result with a larger range. Then, use the first search term set, the first search result, and the user's historical search log to conduct an in-depth analysis, further predict the user's search intention, and construct a second search term set that better meets the user's needs. At the same time, use the second search term set and the scene information corresponding to the second search term set to obtain a rewritten search term that better meets the user's personalized characteristics, so that the rewritten search term is more in line with the user's search intention, thereby meeting the user's personalized search needs, improving search efficiency and search accuracy, and thus improving the user's search experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0013] Figure 1 It is a flowchart of a target search method provided in an embodiment of the present application;

[0014] Figure 2 is a structural schematic diagram of a target search device provided in an embodiment of the present application;

[0015] Figure 3 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0016] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0017] A target search method, device, electronic device, and readable storage medium according to an embodiment of the present application will be described in detail below with reference to the accompanying drawings.

[0018] Figure 1 is a flow chart of a target search method provided in an embodiment of the present application. Figure 1 As shown, the method includes:

[0019] S101, obtaining a search term input by a user;

[0020] S102, constructing a first search term using the search term and the user portrait information, and obtaining a first search result corresponding to the first search term;

[0021] S103, constructing a second search term using the first search term, the first search result, and the user's historical search log;

[0022] S104, determining scene information corresponding to the second search term, obtaining a rewritten search term based on the second search term and the scene information, and performing a search using the rewritten search term to obtain a target search result corresponding to the rewritten search term.

[0023] According to the technical solution provided in the embodiment of the present application, a first search term set is constructed through the search terms and user portrait information input by the user, and a preliminary search is performed based on this, so that a first search result with a larger search range can be obtained. Secondly, the first search term set, the first search result and the user's historical search log are used to perform in-depth analysis to further predict the user's search intention, that is, the first search result and the user's historical search log are used to filter from the first search term set to construct a second search term set that better meets the user's needs, and at the same time, the second search term set and the scene information corresponding to the second search term set are used to obtain rewritten search terms that better meet the user's personalized characteristics, so that the rewritten search terms are more in line with the user's search intention, thereby meeting the user's personalized search needs, improving search efficiency and search accuracy, and thereby improving the user's search experience.

[0024] In some embodiments, a first search term is constructed using search terms and user portrait information, and a first search result corresponding to the first search term is obtained, including: matching the search term with the user portrait information to obtain an intent matching result related to the user's search intent; based on the intent matching result, semantically expanding the search term according to preset semantic rules to obtain the first search term; based on the first search term, obtaining search results that match the first search term, sorting the search results, and selecting a preset number of search results as the first search result based on the sorting result.

[0025] Specifically, by matching the search terms entered by the user with the user portrait information, the user's search intent can be accurately understood. User portrait information includes but is not limited to multi-dimensional information such as the user's age, gender, and hobbies. Through the matching algorithm, the intent matching results related to the search terms and matching the user portrait information can be found. For example, if a user enters the search term "clothes" and the user portrait of the user is: male, 23 years old, likes sports, then the intent matching results can be related keywords such as "sportswear" and "male sportswear".

[0026] It can be understood that according to the intent matching result, the input search term is semantically expanded according to the preset semantic rules. The semantic expansion is specifically expanded using a large language model. The large language model can use deep learning technology to deeply understand and analyze the input search term, and combine the multi-dimensional features in the user portrait information to generate a first search term that is more relevant to the search intent. For example, in the above example, based on the search term "clothes" and the user's portrait information of "male, 23 years old, likes sports", the large language model may expand the search term to more specific and segmented search terms such as "23-year-old male sportswear" and "sports brand men's clothing". At the same time, the large language model can also solve the problem of low search results for long-tail words and improve the search results of the search system.

[0027] It should be noted that the specific semantic expansion operations of the large language model may include but are not limited to synonym replacement, synonym expansion, word form restoration, semantic disambiguation and other operations, which are not specifically limited here to ensure that the revised search term (i.e., the first search term) can more accurately reflect the user's search intent.

[0028] Further, based on the corrected first search term, search results matching the first search term are obtained from a search engine or a related data source. These search results may include various forms of information such as web links, pictures, videos, and commodities. Then, these search results are preliminarily screened to remove content that is irrelevant to the search term or of low quality, and the screened search results are sorted. The basis for sorting may include multiple factors such as the degree of match between the search results and the search term, the authority of the search results, and the like. The sorting algorithm will comprehensively consider these factors, and finally, according to the sorting results, a preset number of search results are selected as the first search results to be displayed to the user, so as to improve search efficiency and reduce the time cost of user screening. This preset number can be adjusted according to actual needs, taking into account the balance between user experience and search efficiency. In this embodiment, the first ten or twenty search results are selected as the first search results.

[0029] According to the technical solution provided in the embodiment of the present application, the search term is matched with the user portrait information to obtain an intent matching result related to the user's search intent; based on the intent matching result, the search term is semantically expanded according to a preset semantic rule to obtain a first search term; based on the first search term, a search result matching the first search term is obtained, the search results are sorted, and a preset number of search results are selected as the first search result based on the sorting result. By matching the search term with the user portrait information, the user's search intent can be understood more accurately. The accuracy of the search is improved, and at the same time, search results that are more in line with their personalized needs are provided to users. The search results are made more in line with the user's actual search needs, improving the user experience.

[0030] In addition, in some embodiments, a second search term is constructed using the first search term, the first search result, and the user's historical search log, including: extracting feature information from the historical search log to obtain historical search features; based on the historical search features, filtering the first search results to obtain a subset of search results related to the historical search features; performing cluster analysis on the search result subset to obtain multiple search clusters; extracting keywords from each search cluster, and combining the keywords with the first search term to obtain multiple candidate search terms; scoring the candidate search terms to obtain scoring results for the multiple candidate search terms, and selecting the candidate search term with the highest score as the second search term based on the scoring results.

[0031] Specifically, in the process of constructing the second search term, historical search features are first extracted from the user's historical search log. The historical search features may include keywords that the user frequently searches for, the search time period, the click-through rate of search results, etc. These historical search features can reflect the user's search habits and interest preferences, and can further understand the user's search intent. Based on the above historical search features, the first search results will be screened, and search results that are irrelevant to the user's historical search or have a low matching degree will be removed, so as to obtain a subset of search results that are highly relevant to the historical search features. This ensures that subsequent analysis and processing are more accurate and efficient.

[0032] Furthermore, the filtered search result subsets are clustered using a cluster analysis algorithm. Cluster analysis is an unsupervised learning method that can classify similar search results into one category to form multiple search clusters. Multiple search clusters can represent different search needs and interests of the user, providing a basis for the subsequent construction of the second search term. After obtaining multiple search clusters, the most representative keywords are extracted from each cluster. These keywords can accurately reflect the theme and content of the cluster, and these keywords are combined with the first search term to generate multiple candidate search terms. The candidate search terms are scored using a scoring algorithm. The scoring algorithm will comprehensively consider multiple factors, such as the relevance of keywords, the click-through rate of search results, etc., to score the candidate search terms. According to the scoring results, the candidate search term with the highest score is selected as the second search term. In this way, it is not only ensured that the second search term meets the user's current search needs, but also fully considers the user's personalized needs and historical search habits.

[0033] Continuing with the above example, for example, the user searches for "male sportswear" for the first time and gets relevant results, then clicks on the search results related to "basketball shoes", and searches for basketball-related content multiple times in the following period of time. By extracting the user's historical search logs, it can be found that the user has a high interest in basketball and sports equipment. Therefore, when constructing the second search term, special attention will be paid to the search results related to "basketball" and "sports equipment". Through cluster analysis, two search clusters are obtained, one cluster is related to "basketball shoes" and the other cluster is related to "basketball sports equipment". From these two clusters, keywords such as "basketball shoes", "basketball uniforms", and "basketball protective gear" are extracted, and these keywords are combined with the first search term "male sportswear" to obtain multiple candidate search terms, such as "male basketball shoes" and "basketball sportswear". Then, these candidate search terms are scored using a scoring algorithm, and the highest-scoring "basketball sportswear" is selected as the second search term.

[0034] According to the technical solution provided in the embodiment of the present application, historical search features are obtained by extracting feature information from historical search logs; based on the historical search features, the first search results are screened to obtain a subset of search results related to the historical search features; cluster analysis is performed on the search result subset to obtain multiple search clusters; keywords are extracted from each search cluster, and the keywords are combined with the first search term to obtain multiple candidate search terms; the candidate search terms are scored to obtain scoring results for the multiple candidate search terms, and the candidate search term with the highest score is selected as the second search term based on the scoring results, so that users can be provided with more accurate and personalized search services.

[0035] In addition, in some embodiments, the scene information includes environmental information and hot spot information, and a rewritten search term is obtained based on the second search term and the scene information, including: using the environmental information to determine environmental keywords; determining hot topic events associated with the search term, and extracting hot keywords from the hot topic events, wherein the similarity between the hot keywords and the second search term is greater than a first preset value; and obtaining a rewritten search term based on the environmental keywords, the hot keywords and the second search term.

[0036] Specifically, in the process of constructing the rewritten search terms, this embodiment fully considers the influence of scene information. Scene information, including at least one of environmental information and hot information, is crucial for a deep understanding of the user's search intent. Environmental information, such as the user's geographical location, weather conditions, etc., can directly affect the user's search needs. Hot information, such as currently popular hot topics, major events, etc., may arouse the user's interest and thus change their search direction.

[0037] It is understandable that environmental keywords are determined using environmental information. For example, if the user is at a basketball court where a Shanxi basketball team is located, and it is currently summer, keywords such as "Shanxi basketball" and "summer sports" can be determined as environmental keywords. These keywords can reflect the user's current environment and help to more accurately understand the user's search intent. In addition, hot topic events associated with the search terms are determined, and hot keywords are extracted from these hot topic events. These hot keywords need to meet the condition that the similarity with the second search term is greater than the first preset value to ensure that they are highly relevant to the user's search intent. For example, if the second search term is "basketball sportswear" and there is currently a hot topic about the Shanxi basketball team winning the championship, keywords such as "Shanxi basketball star", "basketball event", and "winning" can be extracted as hot keywords.

[0038] Finally, based on the environment keywords, hot keywords and the second search terms, rewritten search terms are obtained. Continuing with the above example, keywords such as "Shanxi basketball star" and "win the championship" can be combined with "basketball sportswear" to obtain rewritten search terms such as "Shanxi basketball champion sportswear" and "sportswear of basketball champion star". The rewritten search terms obtained in this way can not only meet the basic search needs of users, but also guide users to discover more information related to the current environment and hot spots, thereby improving user experience.

[0039] According to the technical solution of the embodiment of the present application, the environment information is used to determine the environment keywords; the hot topic events associated with the search words are determined, and the hot keywords are extracted from the hot topic events, wherein the similarity between the hot keywords and the second search words is greater than the first preset value; based on the environment keywords, the hot keywords and the second search words, the rewritten search words are obtained, and the environment keywords, the hot keywords and the second search words are combined to provide users with more accurate search results that are in line with the current scene. This not only improves the search efficiency, but also improves the user's search satisfaction.

[0040] In some embodiments, a rewritten search term is obtained based on environmental keywords, hot keywords and a second search term, including: determining a preset weight allocation strategy corresponding to the user based on user portrait information and historical search logs; determining environmental weights, hotspot weights and search weights based on the preset weight allocation strategy; and weightedly combining environmental keywords, hotspot keywords and the second search term based on the environmental weights, hotspot weights and search weights to obtain a rewritten search term.

[0041] Specifically, in the process of constructing rewritten search terms, user portrait information and historical search logs play an important role in determining the weight distribution of search terms. User portrait information contains the user's personalized characteristics, such as age, gender, interests and hobbies, which helps to understand the user's search preferences and needs. Historical search logs record the user's search history and reflect the user's search habits and behavior patterns. By comprehensively considering user portrait information and historical search logs, a weight distribution strategy that better meets the user's personalized needs can be formulated for the user.

[0042] As an example, for example, for young users, they may pay more attention to current hot topics and popular elements, so the weight of hot keywords can be appropriately increased; while for older users, they may pay more attention to practicality and stability, so the weight of the search term itself should be dominant. In addition, if the user portrait information shows that the user often moves in a specific area, or searches related to the area frequently appear in the historical search log, then when constructing the rewritten search term, the weight of the environmental keyword should be increased accordingly. According to the above-mentioned weight allocation strategy, the environmental weight, hot spot weight and search weight can be determined. Then, according to these weights, the environmental keywords, hot spot keywords and search terms are weighted and combined to obtain the final rewritten search term. For example, if it is determined based on the user portrait and historical search log that a user's search needs are more inclined to practical information related to the second search term, then when constructing the rewritten search term, the weight of the environmental keyword should be dominant, so as to obtain a rewritten search term such as "summer basketball uniform recommendation". If the user's user profile information shows that he is a young male with a strong interest in basketball, and basketball-related search terms frequently appear in historical search logs, then when constructing rewritten search terms, higher weights can be assigned to basketball-related hot keywords. Specifically, for the environmental keyword "Shanxi basketball", the hot keywords "basketball events" and "winning", and the original search term "basketball sportswear", different weights can be assigned to them according to the preset weight allocation strategy. For this user, the weights of "basketball events" and "winning" may be higher than other keywords. At this time, the environmental keywords, hot keywords and original search terms can be weighted and combined according to these weights to obtain rewritten search terms, such as "Shanxi basketball star championship sportswear". Such rewritten search terms not only take into account the user's personalized characteristics and search habits, but also combine the current environmental information and hot information, and can more accurately meet the user's search needs.

[0043] According to the technical solution provided in the embodiment of the present application, a personalized weight allocation strategy is formulated for the user in combination with user portrait information and historical search logs, and the environmental keywords, hot keywords and original search terms are weighted and combined according to these strategies to obtain rewritten search terms. Such rewritten search terms not only take into account the personalized characteristics and search habits of the user, but also combine the current environmental information and hot information, which can provide users with more accurate and personalized search services, improve search efficiency and user search satisfaction, and further enhance user experience.

[0044] In addition, in some embodiments, searching using the rewritten search term to obtain a target search result corresponding to the rewritten search term includes: calculating the similarity between the rewritten expanded term and the third search term, wherein the similarity between the third search term and the search term is greater than a first preset value; if the similarity between the rewritten expanded term and the third search term is greater than a second preset value, searching using the rewritten search term to obtain a target search result corresponding to the rewritten search term, wherein the second preset value is greater than the first preset value.

[0045] Specifically, after the rewritten search terms are obtained, further processing is required to provide more accurate search results. To this end, this embodiment introduces a third search term, which is a vocabulary that is highly related to the search term input by the user, but may not be included in the rewritten search terms. The similarity between the rewritten search terms and the third search terms is calculated to ensure that the search results are not only based on the rewritten search terms themselves, but also take into account the broader context related to the search terms input by the user.

[0046] It can be understood that in the similarity calculation process, the present embodiment pre-sets two preset values: a first preset value and a second preset value. The first preset value is used to preliminarily screen the third search term with a high similarity to the original search term, and the second preset value is used to further screen the rewritten search term with a higher similarity to the third search term. This ensures that the search results obtained are both related to the original search term and highly matched with the rewritten search term. It should be noted that the first preset value and the second preset value can be set according to actual needs, and the specific values ​​of the first preset value and the second preset value are not specifically limited here.

[0047] According to the technical solution provided in the embodiment of the present application, the similarity between the rewritten extension word and the third search word is calculated, wherein the similarity between the third search word and the search word is greater than the first preset value; if the similarity between the rewritten extension word and the third search word is greater than the second preset value, the rewritten search word is used to obtain the target search result corresponding to the rewritten search word, wherein the second preset value is greater than the first preset value, and the third search word is used to further improve the accuracy and relevance of the search results. In this way, more accurate and personalized search services can be provided to users, and the user experience can be improved.

[0048] In addition, in some embodiments, after obtaining the target search result corresponding to the rewritten search term, it also includes: receiving feedback information input by the user, the feedback information is used to represent the user's approval of the search result; if the feedback information indicates that the search result does not meet the user's search intention, the rewritten search term is optimized based on the feedback information.

[0049] Specifically, the user's feedback information can not only directly reflect whether the search results meet their search intent, but also provide a basis for the search engine to improve. Therefore, after obtaining the target search results corresponding to the rewritten search terms, this embodiment specially designs a feedback mechanism to receive user feedback information and optimize the rewritten search terms based on the feedback information.

[0050] It is understandable that the system will display the search results to the user and provide a feedback interface for the user to enter his or her approval of the search results. This feedback interface can be a simple star rating system or a more specific text input box that allows users to describe their feelings and suggestions in detail. When the user submits the feedback information, the system will process the information immediately. If the feedback information indicates that the search results do not meet the user's search intent, the system will start the optimization process of rewriting the search terms. Specifically, the system will analyze the user's feedback content, identify the keywords or topics therein, and adjust the rewritten search terms accordingly. For example, if the user's feedback mentions "I hope to see more information about basketball stars", the system may increase the weight of keywords related to basketball stars in the rewritten search terms, thereby optimizing the search results.

[0051] In addition to directly analyzing user feedback, we can also combine user portrait information and historical search logs to regularly collect and analyze a large amount of user feedback data to discover some common problems and improvement points. For example, if the system finds that a large number of users have reported that "there are too many ads in the search results", then the system may adjust the advertising strategy to reduce the proportion of ads in the search results, thereby improving the user experience.

[0052] According to the technical solution provided in the embodiment of the present application, by receiving feedback information input by the user, the feedback information is used to characterize the user's approval result of the search results; if the feedback information indicates that the search results do not meet the user's search intention, the rewritten search terms are optimized based on the feedback information. The optimization function based on user feedback in this embodiment can not only improve search efficiency, optimize user experience and enhance system intelligence, but also improve the accuracy of search results and promote personalized search.

[0053] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the process of the embodiments of the present application.

[0054] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present application, which will not be described in detail here.

[0055] The following is an embodiment of the device of the present application, which can be used to execute the embodiment of the method of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of the present application.

[0056] Figure 2 Schematic diagram of a target search device provided in an embodiment of the present application. Figure 2 As shown, the device comprises:

[0057] The acquisition module 201 is configured to acquire a search term input by a user;

[0058] A first constructing module 202 is configured to construct a first search term using the search term and the user portrait information, and obtain a first search result corresponding to the first search term;

[0059] A second constructing module 203 is configured to construct a second search term using the first search term, the first search result, and the user's historical search log;

[0060] The search module 204 is configured to determine the scene information corresponding to the second search term, obtain the rewritten search term based on the second search term and the scene information, and perform a search using the rewritten search term to obtain a target search result corresponding to the rewritten search term.

[0061] In some embodiments, the first construction module 202 is also used to match the search term with the user portrait information to obtain an intent matching result related to the user's search intent; based on the intent matching result, the search term is semantically expanded according to preset semantic rules to obtain a first search term; based on the first search term, search results matching the first search term are obtained, the search results are sorted, and a preset number of search results are selected as the first search result based on the sorting result.

[0062] In some embodiments, the second construction module 203 is also used to extract feature information in the historical search log to obtain historical search features; based on the historical search features, the first search results are screened to obtain a subset of search results related to the historical search features; the search result subset is clustered to obtain multiple search clusters; keywords are extracted from each search cluster, and the keywords are combined with the first search term to obtain multiple candidate search terms; the candidate search terms are scored to obtain scoring results for the multiple candidate search terms, and the candidate search term with the highest score is selected as the second search term based on the scoring results.

[0063] In some embodiments, the search module 204 is also used to utilize environmental information to determine environmental keywords; determine hot topic events associated with the search terms, and extract hot keywords from the hot topic events, wherein the similarity between the hot keywords and the second search terms is greater than a preset similarity threshold; and obtain a rewritten search term based on the environmental keywords, the hot keywords and the second search terms.

[0064] In some embodiments, the search module 204 is also used to determine a preset weight allocation strategy corresponding to the user based on user portrait information and historical search logs; determine the environment weight, hotspot weight and search weight based on the preset weight allocation strategy; and perform weighted combination of the environment keywords, hotspot keywords and the second search term based on the environment weight, hotspot weight and search weight to obtain a rewritten search term.

[0065] In some embodiments, the search module 204 is also used to calculate the similarity between the rewritten expansion word and the third search word, wherein the similarity between the third search word and the search word is greater than a first preset value; if the similarity between the rewritten expansion word and the third search word is greater than a second preset value, the rewritten search word is used to obtain a target search result corresponding to the rewritten search word, wherein the second preset value is greater than the first preset value.

[0066] In some embodiments, the search module 204 is also used to receive feedback information input by the user, and the feedback information is used to represent the user's approval of the search results; if the feedback information indicates that the search results do not meet the user's search intent, the rewritten search terms are optimized based on the feedback information.

[0067] Figure 3 Schematic diagram of an electronic device 3 provided in an embodiment of the present application. Figure 3 As shown, the electronic device 3 of this embodiment includes: a processor 301, a memory 302, and a computer program 303 stored in the memory 302 and executable on the processor 301. When the processor 301 executes the computer program 303, the steps in the above-mentioned method embodiments are implemented. Alternatively, when the processor 301 executes the computer program 303, the functions of the modules / units in the above-mentioned device embodiments are implemented.

[0068] The electronic device 3 may be a desktop computer, a notebook, a PDA, a cloud server, or other electronic device. The electronic device 3 may include, but is not limited to, a processor 301 and a memory 302. Those skilled in the art will appreciate that Figure 3 The electronic device 3 is merely an example and does not limit the electronic device 3 , and may include more or less components than those shown in the figure, or different components.

[0069] The processor 301 can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0070] The memory 302 may be an internal storage unit of the electronic device 3, for example, a hard disk or memory of the electronic device 3. The memory 302 may also be an external storage device of the electronic device 3, for example, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 3. The memory 302 may also include both an internal storage unit and an external storage device of the electronic device 3. The memory 302 is used to store computer programs and other programs and data required by the electronic device.

[0071] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units.

[0072] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the present application implements all or part of the process in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. The computer program may include computer program code, which may be in source code form, object code form, executable file or some intermediate form. The readable storage medium may include: any entity or device that can carry computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0073] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A target search method, characterized in that: include: Get the search term entered by the user; Constructing a first search term using the search term and the user portrait information, and obtaining a first search result corresponding to the first search term; constructing a second search term using the first search term, the first search result, and the user's historical search log; Determine the scene information corresponding to the second search word, obtain a rewritten search word based on the second search word and the scene information, and use the rewritten search word to search to obtain a target search result corresponding to the rewritten search word.

2. The method according to claim 1, characterized in that The step of constructing a first search term by using the search term and the user portrait information, and obtaining a first search result corresponding to the first search term, includes: Matching the search term with the user portrait information to obtain an intent matching result related to the user's search intent; Based on the intention matching result, semantically expand the search term according to a preset semantic rule to obtain the first search term; Based on the first search term, search results matching the first search term are obtained, the search results are sorted, and a preset number of the search results are selected as the first search results based on the sorting results.

3. The method according to claim 1, characterized in that The step of constructing a second search term by using the first search term, the first search result, and the user's historical search log includes: Extracting feature information from the historical search log to obtain historical search features; Based on the historical search feature, the first search result is screened to obtain a search result subset related to the historical search feature; Performing cluster analysis on the search result subset to obtain multiple search clusters; Extracting keywords from each search cluster, and combining the keywords with the first search term to obtain a plurality of candidate search terms; The candidate search terms are scored to obtain scoring results of the plurality of candidate search terms, and the candidate search term with the highest score is selected as the second search term according to the scoring results.

4. The method according to claim 1, characterized in that: The scene information includes environment information and hotspot information, and obtaining the rewritten search term based on the second search term and the scene information includes: Using the environmental information, determining environmental keywords; Determine a hot topic event associated with the search term, and extract hot keywords from the hot topic event, wherein the similarity between the hot keywords and the second search term is greater than a preset similarity threshold; The rewritten search term is obtained based on the environmental keyword, the hot keyword and the second search term.

5. The method according to claim 4, characterized in that The obtaining the rewritten search term based on the environment keyword, the hot keyword and the second search term includes: Determine a preset weight allocation strategy corresponding to the user according to the user portrait information and the historical search log; Determine the environment weight, hotspot weight and search weight according to the preset weight allocation strategy; The environment keyword, hotspot keyword and the second search term are weightedly combined according to the environment weight, hotspot weight and search weight to obtain the rewritten search term.

6. The method according to claim 1, characterized in that The step of searching using the rewritten search term to obtain a target search result corresponding to the rewritten search term includes: Calculating the similarity between the rewritten expansion word and a third search word, wherein the similarity between the third search word and the search word is greater than a first preset value; If the similarity between the rewritten extended word and the third search word is greater than a second preset value, the rewritten search word is used for searching to obtain a target search result corresponding to the rewritten search word, wherein the second preset value is greater than the first preset value.

7. The method according to any one of claims 1 to 6, characterized in that: After obtaining the target search result corresponding to the rewritten search term, the method further includes: Receiving feedback information input by a user, wherein the feedback information is used to represent the user's approval result of the search result; If the feedback information indicates that the search result does not meet the user's search intention, the rewritten search term is optimized based on the feedback information.

8. A target search device, characterized in that: include: An acquisition module is configured to acquire a search term input by a user; A first constructing module is configured to construct a first search term using the search term and the user portrait information, and obtain a first search result corresponding to the first search term; A second building module is configured to build a second search term using the first search term, the first search result, and the user's historical search log; The search module is configured to determine the scene information corresponding to the second search term, obtain a rewritten search term based on the second search term and the scene information, and use the rewritten search term to search to obtain a target search result corresponding to the rewritten search term.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Citation Information

Cited By

  • Method and device for determining search input intention, equipment, medium and program product

    CN119271852A

  • Data analysis method and device based on large language model

    CN119397005A