Artificial intelligence-based search result aggregation method, apparatus, and search engine

Abstract

The application discloses a search result aggregation method and device based on artificial intelligence and a search engine. The method comprises the following steps: obtaining a search word; generating a plurality of search results according to the search word; obtaining a plurality of demand dimensions corresponding to the search word; aggregating the plurality of demand dimensions according to the plurality of search results; obtaining answers corresponding to each demand dimension; and aggregating the answers corresponding to the plurality of demand dimensions according to the aggregated demand dimensions to generate an aggregation result. The method reorganizes and displays the search results through multi-level analysis of search demand and results, links to corresponding services accurately while meeting the information demand of users, avoids content redundancy of the search results, reduces the cost of user click searching, and improves the search experience of users.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of search engines, and in particular to a search result aggregation method and device based on artificial intelligence (AI) and a search engine. BACKGROUND

[0002] With the rapid development of computer network technology, search engines have gradually become the main, fastest and most convenient means for people to obtain information. After a user inputs a search term (query) in a search engine, the search engine can return search results for the query to the user.

[0003] At present, a traditional search engine usually gives multiple search results and summaries related to a search term (query) based on the search term, that is, the search engine meets user needs through a method similar to information retrieval. When a user clicks a link corresponding to a search result, the user can view the page content of the search result to find and summarize answers by himself or herself.

[0004] However, there are problems. (1) The search results obtained by the above search method are likely to be homogeneous, which increases the user's click cost. (2) Because the search results are mixed, the readability of the results is poor, which increases the user's search cost. (3) The search demand of the user is not analyzed, so that for relatively complex problems, only one or two results cannot meet the user's search demand, which reduces the search experience. SUMMARY

[0005] The present application aims to at least solve one of the above technical problems.

[0006] To this end, a first object of the present application is to provide a search result aggregation method based on artificial intelligence. The method reorganizes and displays search results through multi-level analysis of search demand and results, accurately links to corresponding services while meeting user information needs, avoids content redundancy of search results, reduces the user's search cost, and improves the user's search experience.

[0007] A second object of the present application is to provide a search result aggregation device based on artificial intelligence.

[0008] A third object of the present application is to provide a search engine.

[0009] To achieve the above object, the search result aggregation method based on artificial intelligence according to the first aspect of the present application comprises the following steps: obtaining a search word; generating a plurality of search results according to the search word; obtaining a plurality of demand dimensions corresponding to the search word; aggregating the plurality of demand dimensions according to the plurality of search results; obtaining an answer corresponding to each demand dimension; and aggregating the answers corresponding to the plurality of demand dimensions according to the aggregated demand dimensions to generate an aggregated result.

[0010] The search result aggregation method based on artificial intelligence according to the present application can obtain a search word, generate a plurality of search results according to the search word, obtain a plurality of demand dimensions corresponding to the search word, aggregate the plurality of demand dimensions according to the plurality of search results, obtain an answer corresponding to each demand dimension, aggregate the answers corresponding to the plurality of demand dimensions according to the aggregated demand dimensions, and generate an aggregated result. The method has at least the following advantages: (1) the aggregation method can remove duplicate results, cross-verify, reorder, etc., reduce repeated clicking behavior of users, and provide aggregated statistical results as a reference to assist user decision-making; and (2) the demand dimensions corresponding to the search word are aggregated, the results of each demand dimension are aggregated, the mixed results are re-integrated, the search results are logically clear and hierarchically distinct, easy to browse, and the user can find the required results more quickly.

[0011] To achieve the above object, the search result aggregation method based on artificial intelligence according to the first aspect of the present application comprises the following steps: obtaining a search word; generating a plurality of search results according to the search word; obtaining a plurality of demand dimensions corresponding to the search word; aggregating the plurality of demand dimensions according to the plurality of search results; obtaining an answer corresponding to each demand dimension; and aggregating the answers corresponding to the plurality of demand dimensions according to the aggregated demand dimensions to generate an aggregated result.

[0012] The search result aggregation device based on artificial intelligence can obtain a search word through the search word obtaining module, generate a plurality of search results according to the search word through the search result obtaining module, obtain a plurality of demand dimensions corresponding to the search word through the demand dimension obtaining module, aggregate the plurality of demand dimensions according to the plurality of search results through the first aggregation module, obtain an answer corresponding to each demand dimension through the answer obtaining module, aggregate the answers corresponding to the plurality of demand dimensions according to the demand dimensions after aggregation through the second aggregation module, and generate an aggregation result.

[0013] To achieve the above object, the search engine of the third aspect of the present application comprises the search result aggregation device based on artificial intelligence of the second aspect of the present application.

[0014] The search engine can obtain a search word through the search word obtaining module in the search result aggregation device, generate a plurality of search results according to the search word through the search result obtaining module, obtain a plurality of demand dimensions corresponding to the search word through the demand dimension obtaining module, aggregate the plurality of demand dimensions according to the plurality of search results through the first aggregation module, obtain an answer corresponding to each demand dimension through the answer obtaining module, and aggregate the answers corresponding to the plurality of demand dimensions according to the demand dimensions after aggregation through the second aggregation module, to generate an aggregation result.

[0015] Additional aspects and advantages of the present application will be made apparent from the following description of embodiments of the present application, taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0016] The above and / or additional aspects and advantages of the present application will become apparent and more readily appreciated from the following description of embodiments, taken in conjunction with the accompanying drawings, in which:

[0017] Figure 1is a flowchart of an artificial intelligence-based search result aggregation method according to one embodiment of the present application;

[0018] Figure 2 is a flowchart of generating a search term demand dimension table according to one embodiment of the present application;

[0019] Figure 3 is a flowchart of generating a search term demand dimension table according to another embodiment of the present application;

[0020] Figure 4 is a flowchart of an artificial intelligence-based search result aggregation method according to one embodiment of the present application;

[0021] Figure 5 is a structural block diagram of an artificial intelligence-based search result aggregation apparatus according to one embodiment of the present application;

[0022] Figure 6 is a structural block diagram of an artificial intelligence-based search result aggregation apparatus according to one embodiment of the present application;

[0023] Figure 7 is a structural block diagram of a first aggregation module according to one embodiment of the present application;

[0024] Figure 8 is a structural block diagram of an artificial intelligence-based search result aggregation apparatus according to another embodiment of the present application. DETAILED DESCRIPTION

[0025] Embodiments of the present application are described in detail below with reference to the attached drawings, which are meant to be exemplary and not limiting.

[0026] Currently, a traditional search engine usually gives relevant N search results (such as URL (Uniform Resoure Locator) addresses) and abstracts based on a search term input by a user, that is, in a manner similar to information retrieval to meet the user's demand. However, this "headache doctor head, foot pain doctor foot" method cannot keep up with the user's usage demand, especially when the problems faced today are becoming more and more complex, and any knowledge has many connotations in addition to its own connotation; the same is true for demand. For example, a user inputs a search term "XXX", in addition to the demand for knowledge encyclopedia, there are also many potential demands such as pictures, videos, audios, news, and the priority of the demand also varies from person to person. On a PC, because the screen is large and can carry a large amount of information, the problem can be solved to some extent by displaying diverse results. However, for mobile terminals, due to the limitation of the screen size and input method of the mobile terminal, the cost of user clicks and inputs is increased. Therefore, the present application provides a search result aggregation method, device and search engine based on artificial intelligence. Specifically, the search result aggregation method, device and search engine based on artificial intelligence of the embodiments of the present application are described below with reference to the accompanying drawings.

[0027] Figure 1 is a flowchart of a search result aggregation method based on artificial intelligence according to an embodiment of the present application. As shown in Figure 1 , the search result aggregation method based on artificial intelligence can include:

[0028] S110, obtaining a search term.

[0029] For example, assuming that the search result aggregation method based on artificial intelligence of the embodiments of the present application is applied to a search engine, the search engine can provide a search term input interface for a user to input a search term (Query) through the input interface. When it is detected that the user inputs a search term through the input interface, the search term can be obtained. The search term can include but is not limited to Chinese, letters, Japanese, Korean, numbers, special characters (such as @, ^, etc.), and the like. In addition, the search term can be a Chinese character text input by the user, and can also be a voice input by the user.

[0030] S120, generating a plurality of search results according to the search term.

[0031] Specifically, after obtaining the search term, a search can be performed according to the search term to search for a plurality of search results related to the search term.

[0032] S130, obtaining a plurality of demand dimensions corresponding to the search term.

[0033] It can be understood that there are many ways to obtain the corresponding multiple demand dimensions according to the search term, for example, the search term can be analyzed online, that is, the demand of the search term is analyzed online to obtain the multiple demand dimensions corresponding to the search term; for example, a search term demand dimension table can also be generated by offline analysis to collect a large number of search terms and corresponding multiple demand dimensions, so that when the search term of the target user is obtained, the multiple demand dimensions corresponding to the search term are queried in the pre-generated search term demand dimension table.

[0034] As an example, the search term demand dimension table can be obtained first, and then the search term can be searched from the search term demand dimension table to obtain the multiple demand dimensions corresponding to the search term.

[0035] It can be understood that the above-mentioned search term demand dimension table can be generated in many ways, for example, the demand dimension can be mined by the user's active change behavior in the user session information; for example, the demand dimension can be mined based on Title and URL information:

[0036] As an example, as shown in the Figure 2 The search term demand dimension table can be pre-generated by the following steps:

[0037] S210, obtaining user session information of multiple users.

[0038] Because there is a large amount of data on the Internet, the user session information generated when a large number of sample users use the search engine can be obtained by using this feature of the Internet. It can be understood that the user session information can include but is not limited to the search term input by the sample user.

[0039] S220, demand mining according to the user active change behavior information in the user session information to generate multiple demand dimensions.

[0040] It can be understood that the user session information can contain multiple search terms input by the user, and there is a correlation between the multiple search terms. Therefore, the multiple search terms input by the user can be summarized according to the correlation between the multiple search terms, and then the demand mining of the active change behavior of the user for these search terms with correlation can be performed to obtain multiple demand dimensions. For example, taking user session information={XXX, XXX movie, XXX movie download} as an example, it can be seen that the search terms "XXX", "XXX movie" and "XXX movie download" can be contained in the user session information, and there is a correlation between the three search terms. Among them, "movie" is a demand dimension of "XXX", and "download" is a demand dimension of "XXX movie". Therefore, the demand dimension can be mined by the active change behavior of the user in the user session information.

[0041] S230, establishing a corresponding relationship between the search term in the user session information and the plurality of demand dimensions.

[0042] S240, generating a search term-demand dimension correspondence table according to the plurality of demand dimensions, the search term in the user session information, and the corresponding relationship.

[0043] That is, the plurality of demand dimensions and the search term in the user session information can be summarized according to the above-mentioned corresponding relationship to obtain the search term-demand dimension correspondence table.

[0044] As another example, as shown in Figure 3 The search term-demand dimension correspondence table is pre-generated by the following steps:

[0045] S310, obtaining user session information of a plurality of users.

[0046] S320, mining a set of potential demand words according to the syntax structure of the search term in the user session information.

[0047] S330, mining and supplementing the set of potential demand words according to the site information corresponding to the url in the search result corresponding to the search term in the user session information.

[0048] S340, screening and determining a plurality of demand dimensions according to the statistical characteristics of the set of potential demand words.

[0049] S350, establishing a corresponding relationship between the search term in the user session information and the plurality of demand dimensions.

[0050] S360, generating a search term-demand dimension correspondence table according to the plurality of demand dimensions, the search term in the user session information, and the corresponding relationship.

[0051] S140, aggregating the plurality of demand dimensions according to the plurality of search results.

[0052] Specifically, in the embodiments of the present application, the specific implementation process of aggregating the plurality of demand dimensions according to the plurality of search results can be as follows: the syntax structure of the title and the abstract in the plurality of search results can be analyzed, and the site information and the webpage type corresponding to the url in the plurality of search results can be analyzed, and finally, the plurality of demand dimensions can be aggregated according to the analyzed title syntax structure, abstract syntax structure, site information, and webpage type.

[0053] More specifically, each of the plurality of search results includes title, abstract and url information, after which, the syntax analysis result of the title, the syntax analysis result of the abstract, the site information of the url, and the webpage type corresponding to the url can be extracted according to these information respectively, and then, according to these analysis results, the plurality of demand dimensions corresponding to the search term are aggregated to aggregate the same or similar demands, for example, the demands with the same meaning of "what is the matter" and "reason", the demands with the same meaning of "what to do" and "how to solve", the demands with the same meaning of "where to repair" and "fault repair", etc.

[0054] In S150, the answers corresponding to each demand dimension are obtained, and the answers corresponding to the plurality of demand dimensions are aggregated according to the aggregated demand dimensions to generate an aggregated result.

[0055] Specifically, the answer content corresponding to each demand dimension is obtained from the plurality of search results, after which the answers corresponding to each demand dimension are mapped to the aggregated demand dimensions, and the answers mapped to the same demand dimension are aggregated with each other to generate an aggregated result. For example, taking the search term "mobile phone automatically restarts" as an example, under the "fault reason" demand dimension, the search result content will mention many similar opinions such as "software incompatibility" and "mobile phone hardware problem", and the answer content under the "fault reason" demand dimension is aggregated by content aggregation to obtain the result aggregation content under the demand dimension. It can be understood that in order to avoid redundancy of the result content, after the answers corresponding to the plurality of demand dimensions are aggregated according to the aggregated demand dimensions, the aggregated result can be normalized to avoid the same content from appearing repeatedly.

[0056] More specifically, since each of the plurality of search results includes title, abstract and url information, the syntax analysis result of the title, the syntax analysis result of the abstract, the site information of the url, and the webpage type corresponding to the url can be extracted according to these information respectively, and then, according to these analysis results, a machine learning model is established based on a large amount of user conversation information, each search result is mapped to a certain demand dimension (such as the aggregated demand dimension described above), and finally, the results mapped to the same demand dimension are aggregated with each other to obtain an aggregated result.

[0057] Specifically, the aggregation of the answers corresponding to the plurality of demand dimensions according to the demand dimension after the aggregation is a typical machine learning process in artificial intelligence. First, according to the title, abstract and url information of the search results, a feature set including but not limited to semantic features, syntactic structure features, demand word features, url site type features and the like is extracted, and those skilled in the art can adjust the feature set according to the existing user session data, the completeness of the semantic and page classification tools. Then, a plurality of classifiers are trained by extracting the feature set from a large amount of user session information with demand dimension annotation, and the mapping relationship between the feature set and the demand dimension set is learned. The plurality of classifiers used in this step can be a traditional machine learning model such as a support vector machine, an integrated machine learning model such as a random forest, or a complex machine learning model such as a deep neural network. Those skilled in the art can select the appropriate model through trial and comparison according to the data volume, the number of features and the performance of the computer device. Finally, after the search word obtains a plurality of search results, the features of the plurality of search results are extracted and mapped through the plurality of classifiers, and the demand dimension corresponding to each search result is obtained, and the aggregation is completed according to whether the demand dimensions are the same.

[0058] For example, taking the search word "XXX" as an example, there are two results with titles "XXX latest movie online viewing" and "XXX classic movie viewing" in the plurality of search results. Through syntactic analysis of the title, it can be known that the demand word of the two results is "viewing", and the url site is a video online website. The trained machine learning model can map the two results to the "video" demand dimension at the same time through these features. Finally, the two results are aggregated together because they are mapped to the same demand dimension. It can be understood that in the manner shown by the example, the plurality of search results can be aggregated, and one or more search results can be associated with each demand dimension associated with the search word.

[0059] Therefore, through multi-level analysis of search demand and results, the search results are reorganized and displayed, which can accurately link to services while meeting the information needs of users, such as directly providing official repair merchant addresses under the "repair merchant" dimension, or even customer service platforms. It can be understood that different demand dimensions can provide users with service content corresponding to the demand dimension, such as question and answer content, purchase information content, repair service content, etc.

[0060] In order for those skilled in the art to more clearly understand the present application, the following examples are provided.

[0061] For example, taking the search word "XXX" as an example, after obtaining the search word "XXX", a plurality of search results corresponding to the search word can be generated, and a plurality of demand dimensions corresponding to the search word can be queried from the search word demand dimension correspondence table, such as the "knowledge encyclopedia" demand dimension, the "star itinerary" demand dimension, the "fan attention" video dimension, the "star microblog" dimension, the "picture" dimension, and the "news" dimension. Then, the plurality of demand dimensions are aggregated according to the plurality of search results obtained above, that is, the plurality of demand dimensions are classified according to the search results, that is, demand dimensions with the same meaning are classified into a category, so as to complete the aggregation of the plurality of demand dimensions. Then, the answer content corresponding to each demand dimension (i.e., the demand dimension obtained first) can be obtained from the plurality of search results, and finally, the answers corresponding to each demand dimension are aggregated according to the aggregated demand dimensions to obtain an aggregated result. The aggregated result contains the above demand dimensions and the result content corresponding to each demand dimension.

[0062] For example, taking the search word "XXX" as an example, after obtaining the search word "XXX", a plurality of search results corresponding to the search word can be generated, and a plurality of demand dimensions corresponding to the search word can be queried from the search word demand dimension correspondence table, such as the "knowledge encyclopedia" demand dimension, the "star itinerary" demand dimension, the "fan attention" video dimension, the "star microblog" dimension, the "picture" dimension, and the "news" dimension. Then, the plurality of demand dimensions are aggregated according to the plurality of search results obtained above, that is, the plurality of demand dimensions are classified according to the search results, that is, demand dimensions with the same meaning are classified into a category, so as to complete the aggregation of the plurality of demand dimensions. Then, the answer content corresponding to each demand dimension (i.e., the demand dimension obtained first) can be obtained from the plurality of search results, and finally, the answers corresponding to each demand dimension are aggregated according to the aggregated demand dimensions to obtain an aggregated result. The aggregated result contains the above demand dimensions and the result content corresponding to each demand dimension.

[0063] To sum up, the search result aggregation method based on artificial intelligence in the embodiment of the application aggregates the search request of the user in the demand dimension first, classifies and arranges the search results through multi-dimensional analysis of the results for a single demand problem, and extracts dimensional features to achieve accurate satisfaction of information and accurate provision of services. For a multi-demand problem, the demand is classified and aggregated first, and then is displayed in layers, so that the user can find the required without reinitiating a new search. Thus, while satisfying the single demand problem, the intelligent recommendation of related problems and related results provides the user with an immersive experience. For example, the user inputs "mobile phone automatic restart", and through demand analysis and aggregation, "fault cause", "solution", "repair business" and other demands can be obtained. For "solution", it can be divided into "mobile phone model 1" and "mobile phone model 2" and other dimensions. For a specified dimension such as "mobile phone model 2 solution", through result aggregation and induction, "battery end pad paper", "press the battery seat gold finger", "plug in the charger" and other actual solutions can be summarized, so as to solve the actual problem of the user in multiple levels and multiple dimensions.

[0064] The search result aggregation method based on artificial intelligence in the embodiment of the application can first acquire a search word, then generate a plurality of search results according to the search word, acquire a plurality of demand dimensions corresponding to the search word, aggregate the plurality of demand dimensions according to the plurality of search results, acquire answers corresponding to each demand dimension, and aggregate the answers corresponding to the plurality of demand dimensions according to the demand dimensions after aggregation, to generate an aggregated result. The method has at least the following advantages: (1) through aggregation, the results can be de-duplicated, cross-verified, reordered, etc., reducing the repeated clicking behavior of the user, and providing aggregated statistical results as a reference to assist the user in decision-making; (2) through aggregation of the demand dimensions corresponding to the search word and aggregation of the result content of each demand dimension, the mixed results are reorganized, so that the search results are logically clear and hierarchical, easy to browse, and thus the user can find the required more quickly.

[0065] Figure 4 is a flowchart of the search result aggregation method based on artificial intelligence according to one specific embodiment of the application.

[0066] In order to improve the search experience of the user and make the user find the required in an immersive browsing experience, in the embodiment of the application, related results of other dimensions can also be given through guidance, stimulation and other means. Specifically, as shown in Figure 4 The search result aggregation method based on artificial intelligence can include:

[0067] S410, acquiring a search word.

[0068] S420, generating a plurality of search results according to the search word.

[0069] S430, obtaining a plurality of demand dimensions corresponding to the search term according to the search term.

[0070] As an example, a search term demand dimension correspondence table can be obtained first, and then the search term demand dimension correspondence table can be searched according to the search term to obtain a plurality of demand dimensions corresponding to the search term.

[0071] It can be understood that there are many ways to generate the search term demand dimension correspondence table described above, for example, demand dimensions can be mined through user active change behaviors in user session information; for another example, demand dimensions can be mined based on Title and URL information:

[0072] As an example, as shown in Figure 2 The search term demand dimension correspondence table is pre-generated by the following steps:

[0073] S210, obtaining user session information of a plurality of users.

[0074] S220, demand mining according to user active change behavior information in the user session information to generate a plurality of demand dimensions.

[0075] S230, establishing a corresponding relationship between the search term in the user session information and the plurality of demand dimensions.

[0076] S240, generating a search term demand dimension correspondence table according to the plurality of demand dimensions, the search term in the user session information, and the corresponding relationship.

[0077] As another example, as shown in Figure 3 The search term demand dimension correspondence table is pre-generated by the following steps:

[0078] S310, obtaining user session information of a plurality of users.

[0079] S320, mining a set of potential demand words according to the syntactic structure of the search term in the user session information.

[0080] S330, mining and supplementing the set of potential demand words according to the site information corresponding to the URL in the search result corresponding to the search term in the user session information.

[0081] S340, screening and determining a plurality of demand dimensions according to the statistical characteristics of the set of potential demand words.

[0082] S350, establishing a corresponding relationship between the search term in the user session information and the plurality of demand dimensions.

[0083] S360, generating a search term demand dimension correspondence table according to the plurality of demand dimensions, the search term in the user session information, and the corresponding relationship.

[0084] S440, aggregating the plurality of demand dimensions according to the plurality of search results.

[0085] Specifically, in the embodiment of the present application, the specific implementation process of aggregating the plurality of demand dimensions according to the plurality of search results can be as follows: the syntax structure of the title and the abstract in the plurality of search results can be analyzed first, and the site information and the web page type corresponding to the url in the plurality of search results can be analyzed, and finally, the plurality of demand dimensions can be aggregated according to the title syntax structure, the abstract syntax structure, the site information and the web page type obtained by the analysis.

[0086] S450, obtaining the answer corresponding to each demand dimension, and aggregating the answers corresponding to the plurality of demand dimensions according to the aggregated demand dimensions to generate an aggregated result.

[0087] Specifically, in an embodiment of the present application, the answer content corresponding to each demand dimension can be obtained from the plurality of search results, and then the answer corresponding to each demand dimension is mapped to the aggregated demand dimension, and the answers mapped to the same demand dimension are aggregated with each other to generate an aggregated result.

[0088] S460, determining the demand dimensions not covered in the aggregated result, and presenting the demand dimensions not covered in the aggregated result in the form of related search words on the search result page.

[0089] For example, after obtaining the aggregated result, the demand dimensions not covered in the aggregated result can be presented in the form of related search words on the search result page based on the above search word demand dimension correspondence table, so as to provide the user with heuristic guidance. In this way, when the user clicks on the related search word, a new search demand can be triggered, and the step S410 is returned, thereby achieving the purpose of fully meeting the user's demand.

[0090] The search result aggregation method based on artificial intelligence in the embodiment of the present application can determine the demand dimensions not covered in the aggregated result after obtaining the aggregated result, and present the demand dimensions not covered in the aggregated result in the form of related search words on the search result page. That is, the demand dimensions not covered in the aggregated result are presented in the form of related search words on the search result page to provide the user with heuristic guidance. In this way, through guidance, stimulation and other means, related results of other dimensions are given to the user to find what is needed in an immersive browsing experience.

[0091] Corresponding to the search result aggregation method based on artificial intelligence provided by the above several embodiments, an embodiment of the present application further provides a search result aggregation device based on artificial intelligence. Since the search result aggregation device based on artificial intelligence provided by the embodiment of the present application corresponds to the search result aggregation method based on artificial intelligence provided by the above several embodiments, the implementation of the foregoing search result aggregation method based on artificial intelligence is also applicable to the search result aggregation device based on artificial intelligence provided by the present embodiment. In the present embodiment, it will not be described in detail. Figure 5 is a structural block diagram of a search result aggregation device based on artificial intelligence according to an embodiment of the present application. As shown in Figure 5 , the search result aggregation device based on artificial intelligence can include a search term acquisition module 100, a search result acquisition module 200, a demand dimension acquisition module 300, a first aggregation module 400, an answer acquisition module 500, and a second aggregation module 600.

[0092] Specifically, the search term acquisition module 100 can be used to acquire a search term.

[0093] The search result acquisition module 200 can be used to generate a plurality of search results according to the search term.

[0094] The demand dimension acquisition module 300 can be used to acquire a plurality of corresponding demand dimensions according to the search term. Specifically, in an embodiment of the present application, the demand dimension acquisition module 300 can first acquire a pre-generated search term demand dimension correspondence table, and then can search the search term demand dimension correspondence table according to the search term to acquire a plurality of demand dimensions corresponding to the search term.

[0095] Further, in an embodiment of the present application, as shown in Figure 6 , the search result aggregation device based on artificial intelligence can further include a pre-processing module 700. The pre-processing module 700 can be used to acquire user session information of a plurality of users, and perform demand mining according to user active change behavior information in the user session information to generate a plurality of demand dimensions, and establish a correspondence between the search term in the user session information and the plurality of demand dimensions, and generate a search term demand dimension correspondence table according to the plurality of demand dimensions, the search term in the user session information and the correspondence.

[0096] In another embodiment of the present application, the preprocessing module 700 can also be configured to obtain user session information of a plurality of users, mine a set of potential demand words according to the syntactic structure of search words in the user session information, mine and supplement the set of potential demand words according to the site information corresponding to the url in the search result corresponding to the search words in the user session information, filter and determine a plurality of demand dimensions according to the statistical characteristics of the set of potential demand words, establish a corresponding relationship between the search words in the user session information and the plurality of demand dimensions, and generate a search word-demand dimension correspondence table according to the plurality of demand dimensions, the search words in the user session information and the corresponding relationship.

[0097] The first aggregation module 400 can be configured to aggregate the plurality of demand dimensions according to the plurality of search results. Specifically, in an embodiment of the present application, as shown in Figure 7 the first aggregation module 400 can include a first analysis unit 410, a second analysis unit 420 and an aggregation unit 430. The first analysis unit 410 can be configured to analyze the syntactic structure of the title and the abstract in the plurality of search results. The second analysis unit 420 can be configured to analyze the site information and the webpage type corresponding to the url in the plurality of search results. The aggregation unit 430 is configured to aggregate the plurality of demand dimensions according to the title syntactic structure, the abstract syntactic structure, the site information and the webpage type obtained by analysis.

[0098] The answer obtaining module 500 can be configured to obtain the answer corresponding to each demand dimension. Specifically, in an embodiment of the present application, the answer obtaining module 500 can obtain the answer corresponding to each demand dimension from the plurality of search results.

[0099] The second aggregation module 600 can be configured to aggregate the answers corresponding to the plurality of demand dimensions according to the demand dimensions after aggregation, to generate an aggregation result. Specifically, in an embodiment of the present application, the second aggregation module 600 can map the answer corresponding to each demand dimension to the demand dimension after aggregation, and aggregate the answers mapped to the same demand dimension with each other to generate an aggregation result.

[0100] Further, in an embodiment of the present application, as shown in Figure 8 the artificial intelligence-based search result aggregation apparatus can further include a determination module 800 and a presentation module 900. The determination module 800 can be configured to determine the demand dimensions not covered in the aggregation result. The presentation module 900 can be configured to present the demand dimensions not covered in the aggregation result in the form of related search words on a search result page.

[0101] The search result aggregation device based on artificial intelligence can obtain a search word through the search word obtaining module, generate a plurality of search results according to the search word through the search result obtaining module, obtain a plurality of demand dimensions corresponding to the search word through the demand dimension obtaining module, aggregate the plurality of demand dimensions according to the plurality of search results through the first aggregation module, obtain an answer corresponding to each demand dimension through the answer obtaining module, and aggregate the answers corresponding to the plurality of demand dimensions according to the demand dimensions after aggregation through the second aggregation module, to generate an aggregation result. At least the following advantages are obtained: (1) the aggregation method is used to remove, cross-verify, reorder, etc., to reduce repeated clicking behavior of a user, and the aggregation statistical result can be provided as a reference to assist user decision-making; and (2) the demand dimensions corresponding to the search word are aggregated, and the result content of each demand dimension is aggregated, that is, the mixed results are re-integrated, so that the search result is logically clear and hierarchical, is convenient to browse, and can help a user to find a required result more quickly.

[0102] To achieve the above-mentioned embodiments, the application further provides a search engine, which can include the search result aggregation device based on artificial intelligence according to any one of the above-mentioned embodiments of the application.

[0103] The search engine can obtain a search word through the search word obtaining module in the search result aggregation device, generate a plurality of search results according to the search word through the search result obtaining module, obtain a plurality of demand dimensions corresponding to the search word through the demand dimension obtaining module, aggregate the plurality of demand dimensions according to the plurality of search results through the first aggregation module, obtain an answer corresponding to each demand dimension through the answer obtaining module, and aggregate the answers corresponding to the plurality of demand dimensions according to the demand dimensions after aggregation through the second aggregation module, to generate an aggregation result. At least the following advantages are obtained: (1) the aggregation method is used to remove, cross-verify, reorder, etc., to reduce repeated clicking behavior of a user, and the aggregation statistical result can be provided as a reference to assist user decision-making; and (2) the demand dimensions corresponding to the search word are aggregated, and the result content of each demand dimension is aggregated, that is, the mixed results are re-integrated, so that the search result is logically clear and hierarchical, is convenient to browse, and can help a user to find a required result more quickly.

[0104] In the description of the application, it should be understood that the terms "first", "second" are only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first", "second" can explicitly or implicitly include at least one of the features. In the description of the application, the meaning of "a plurality of" is at least two, for example, two, three, etc., unless otherwise specifically limited.

[0105] In the description of the application, reference has been made to descriptive terms such as "one embodiment", "some embodiments", "an example", "a specific example" or "some examples" etc. Such terminology means that a particular feature or characteristic described in connection with such embodiments or examples is included in at least one embodiment or example of the application. The illustrative discussion of such terms are not meant to be limiting in that the same features or characteristics need not be present in all embodiments or examples of the application. Nor are the features or characteristics described in connection with one embodiment or example necessarily excluded from inclusion in another embodiment or example of the application. Furthermore, the description herein of one or more embodiments or examples of the application, including the description of the best mode of practicing such embodiments or examples, is not necessarily meant to exclude features or characteristics of the application from use in at least one other embodiment or example of the application. One of ordinary skill in the art will recognize that many of the embodiments or examples of the application can be implemented using one or more computer program products that can be traded as trade articles.

[0106] Any process or method described in flow charts or otherwise described herein can be understood as representing a module, segment, or portion of code that includes one or more executable instructions for implementing specific logical functions or steps, and the various embodiments of the application contemplate that additional implementation can be implemented in which the order of steps can be changed, including that the steps can be performed in substantially simultaneous fashion or in reverse order, depending on the functionality involved, as will be understood by those having ordinary skill in the art to which the embodiments of the application pertain.

[0107] The logic and / or steps represented in flow charts or otherwise described herein, for example, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions, or a combination thereof. For purposes of this specification, a "computer-readable medium" can be any apparatus that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. The computer-readable medium can be a computer- readable storage medium or a computer-readable communication medium. The computer- readable storage medium can be, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a propagation medium. The computer-readable communication medium can be, for example, but is not limited to, a carrier wave or a computer readable medium including R.F., N.R.F., line, wireless, cable, or other communication medium. The computer-readable medium can also be, for example, but is not limited to, one or more of the following: a portable computer diskette (magnetic), a compact disc (CD) (optical), a read-only memory (ROM) (electronic), a random access memory (RAM) (electronic), or a flash memory (electronic). Note that the computer-readable medium can even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example via optical scanning of the paper or other medium, then compiled, interpreted, or otherwise processed in a suitable manner, if necessary, and then stored in a computer memory.

[0108] It should be understood that each part of the present application can be realized by hardware, software, firmware or their combination. In the above-mentioned embodiments, a plurality of steps or methods can be realized by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if realized by hardware, and as in another embodiment, any one or their combination of the following technologies known in the art can be used: discrete logic circuit with logic gate circuit for implementing logic function on data signal, application specific integrated circuit with suitable combination logic gate circuit, programmable gate array (PGA), field programmable gate array (FPGA) and the like.

[0109] Those skilled in the art of the present technology can understand that all or part of the steps carried out by the above-mentioned embodiment method can be completed by a program instructing the relevant hardware, and the program can be stored in a computer readable storage medium, and when the program is executed, it includes one of the steps of the method embodiment or a combination thereof.

[0110] In addition, each functional unit in each embodiment of the present application can be integrated in one processing module, or each unit can exist physically alone, or two or more units can be integrated in one module. The above-mentioned integrated module can be realized in the form of hardware or in the form of software functional module. The integrated module, if realized in the form of software functional module and sold or used as an independent product, can also be stored in a computer readable storage medium.

[0111] The above-mentioned storage medium can be read-only memory, magnetic disk or optical disk, etc. Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. An artificial intelligence-based search result aggregation method, characterized by, The method comprises the following steps: obtaining a search word; generating a plurality of search results according to the search word; obtaining a plurality of demand dimensions corresponding to the search word; the plurality of demand dimensions of the search word are obtained through demand mining based on active change behaviors of a user in a user session information with respect to the search word having relevance; aggregating the plurality of demand dimensions according to the plurality of search results; the aggregating the plurality of demand dimensions according to the plurality of search results comprises: analyzing the syntax structure of the title and the abstract in the plurality of search results; analyzing the site information and the webpage type corresponding to the url in the plurality of search results; aggregating the plurality of demand dimensions according to the analyzed title syntax structure, abstract syntax structure, site information and webpage type; obtaining answers corresponding to each demand dimension, and aggregating the answers corresponding to the plurality of demand dimensions according to the aggregated demand dimensions to generate an aggregated result; determining demand dimensions not covered in the aggregated result, and presenting the demand dimensions not covered in the aggregated result in the form of a related search word on a search result page to provide the user with heuristic guidance. 2.The artificial intelligence-based search result aggregation method of claim 1, wherein, the obtaining a plurality of demand dimensions corresponding to the search word comprises: obtaining a pre-generated search word demand dimension correspondence table; searching the search word demand dimension correspondence table according to the search word to obtain a plurality of demand dimensions corresponding to the search word. 3.The artificial intelligence-based search result aggregation method of claim 2, wherein, the search word demand dimension correspondence table is pre-generated through the following steps: obtaining user session information of a plurality of users; generating a plurality of demand dimensions according to user active change behavior information in the user session information; establishing a corresponding relationship between the search word in the user session information and the plurality of demand dimensions; generating the search word demand dimension correspondence table according to the plurality of demand dimensions, the search word in the user session information and the corresponding relationship. 4.The artificial intelligence-based search result aggregation method of claim 2, wherein, the search word demand dimension correspondence table is pre-generated through the following steps: obtaining user session information of a plurality of users; mining a potential demand word set according to the syntax structure of the search word in the user session information; mining and supplementing the potential demand word set according to the site information corresponding to the url in the search result corresponding to the search word in the user session information; screening and determining a plurality of demand dimensions according to the statistical characteristics of the potential demand word set; establishing a corresponding relationship between the search word in the user session information and the plurality of demand dimensions; generating the search word demand dimension correspondence table according to the plurality of demand dimensions, the search word in the user session information and the corresponding relationship. 5.The artificial intelligence-based search result aggregation method of claim 1, wherein, the obtaining answers corresponding to each demand dimension, and aggregating the answers corresponding to the plurality of demand dimensions according to the aggregated demand dimensions to generate an aggregated result comprises: obtaining answers corresponding to each demand dimension from the plurality of search results; mapping the answers corresponding to each demand dimension to the aggregated demand dimensions, and aggregating the answers mapped to the same demand dimension to generate an aggregated result.

6. An artificial intelligence-based search result aggregation apparatus, characterized by comprising: comprises: a search word obtaining module for obtaining a search word; The search result obtaining module is configured to generate a plurality of search results according to the search term; The demand dimension obtaining module is configured to obtain a plurality of demand dimensions corresponding to the search term according to the search term; the plurality of demand dimensions of the search term are obtained based on demand mining of active change behaviors of a user with respect to the search term having relevance in user session information; The first aggregation module is configured to aggregate the plurality of demand dimensions according to the plurality of search results; wherein the first aggregation module comprises: The first analysis unit is configured to analyze the syntax structure of the title and the abstract in the plurality of search results; The second analysis unit is configured to analyze the site information and the webpage type corresponding to the URL in the plurality of search results; The aggregation unit is configured to aggregate the plurality of demand dimensions according to the title syntax structure, the abstract syntax structure, the site information, and the webpage type obtained by analysis; The answer obtaining module is configured to obtain an answer corresponding to each demand dimension; The second aggregation module is configured to aggregate the answers corresponding to the plurality of demand dimensions according to the demand dimensions after aggregation, to generate an aggregation result; The determination module is configured to determine a demand dimension not covered in the aggregation result; The presentation module is configured to present the demand dimension not covered in the aggregation result in the form of a related search term on a search result page to provide the user with heuristic guidance.

7. The artificial intelligence-based search result aggregation apparatus of claim 6, wherein, The demand dimension obtaining module is specifically configured to: Obtain a pre-generated search term demand dimension correspondence table; According to the search term, search in the search term demand dimension correspondence table to obtain a plurality of demand dimensions corresponding to the search term.

8. The artificial intelligence-based search result aggregation apparatus of claim 7, wherein, Further comprising: The pre-processing module is configured to obtain user session information of a plurality of users, and perform demand mining according to active change behavior information of the users in the user session information, to generate a plurality of demand dimensions, and establish a corresponding relationship between search terms in the user session information and the plurality of demand dimensions, and generate the search term demand dimension correspondence table according to the plurality of demand dimensions, the search terms in the user session information, and the corresponding relationship. 9.The artificial intelligence-based search result aggregation apparatus of claim 8, wherein, The pre-processing module is further configured to obtain user session information of a plurality of users, and mine a set of potential demand terms according to the syntax structure of the search terms in the user session information, and mine and supplement the set of potential demand terms according to the site information corresponding to the URL in the search results corresponding to the search terms in the user session information, and screen and determine a plurality of demand dimensions according to the statistical characteristics of the set of potential demand terms, and establish a corresponding relationship between search terms in the user session information and the plurality of demand dimensions, and generate the search term demand dimension correspondence table according to the plurality of demand dimensions, the search terms in the user session information, and the corresponding relationship.

10. The search result aggregation device based on artificial intelligence according to claim 6, wherein The answer obtaining module is specifically configured to obtain an answer corresponding to each demand dimension from the plurality of search results; The second aggregation module is specifically configured to map the answers corresponding to each requirement dimension to a requirement dimension after aggregation, and aggregate the answers mapped to the same requirement dimension with each other to generate an aggregation result.

11. A search engine, characterized by, Comprise: The artificial intelligence-based search result aggregation apparatus according to any one of claims 6 to 10.

Images