Searching method and system, electronic equipment, storage medium and program product

Abstract

The embodiment of the invention provides a search method and system, electronic equipment, a storage medium and a program product. According to the technical scheme provided by some embodiments of the invention, the search results are processed and the reference information for assisting the user is output by utilizing the powerful capability of the preset model, and the reference information can be navigation information obtained by sorting and summarizing the search results and can also be recommended information selected from the search results and recommended to the user; according to the embodiment of the invention, the search result obtained based on keyword search and the reference information output by the preset model are displayed on the interface together, so that a user can see the search result and the reference information on the same interface, and the user can conveniently know the conditions of the search results, such as which types or result items worthy of attention exist in the search results. Therefore, according to the scheme of the embodiment of the invention, the search depth and width of the user can be expanded, the user is helped to know the search result, the reading cost is reduced, and the decision-making efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular, to a search method, system, electronic device, storage medium, and program product. Background Art

[0002] When a user has a search need, they can enter keywords through the browser web page or the search page of an application (APP). The search engine can search for search results based on the keywords and present the search results on the web page or APP page. In most cases, the search results are very diverse. For example, in a product search, the products found span different brands, models, etc., making it difficult for the user to know how to choose.

[0003] Most users will simply browse through the results and then click on the page of a product that they may be interested in based on their intuition to view product details, parameters, etc. In this way, the user may miss some products, and the missed products may be ones that the user is truly interested in. Summary of the Invention

[0004] To address the problems in the existing technology, this application provides a search method, system, electronic device, storage medium, and program product in multiple aspects.

[0005] In the first embodiment of this application, a search method is provided. The method includes:

[0006] Respond to a search event and obtain search results related to the keywords;

[0007] Obtain reference information output by a preset model based on the search results to assist the user in understanding the search results;

[0008] On the interface, display the search results and the reference information.

[0009] In the second embodiment of this application, a search method is provided. The method includes:

[0010] Respond to a search event and obtain search results related to the keywords, where the search results are obtained by searching based on a search intent related to the keywords, and the search intent is generated by a first preset model based on the keywords;

[0011] Obtain reference information output by a second preset model based on the search results;

[0012] On the interface, display the search results and the reference information;

[0013] Wherein, the first preset model and the second preset model are the same model or two different models; the reference information is navigation information for assisting the user to understand the search results or recommendation information; the recommendation information includes at least one recommended result item selected by the second preset model from the search results and the recommended text generated for the recommended result item.

[0014] In the third embodiment of the present application, a search method is provided. The method includes:

[0015] Responding to a search event, obtaining a first search result related to the keyword retrieved by a search engine;

[0016] Obtaining a second search result related to the keyword, where the second search result is retrieved by the search engine based on a search intent related to the keyword, and the search intent is generated by a first preset model based on the keyword;

[0017] Obtaining first reference information output by a second preset model based on the second search result;

[0018] On the interface, displaying the first search result and the first reference information;

[0019] Wherein, the first preset model and the second preset model are the same model or two different models; the first reference information is first navigation information for assisting the user to understand the second search result, or first recommendation information; the first recommendation information includes at least one recommended result item selected by the second preset model from the second search results and the recommended text generated for the recommended result item.

[0020] In the fourth embodiment of the present application, a search method is provided. The method includes:

[0021] Obtaining a keyword;

[0022] Searching to obtain search results related to the keyword;

[0023] Using a preset model to process the search results to obtain reference information for assisting the user to understand the search results;

[0024] Sending the search results and the reference information to a client for display on the client interface.

[0025] In the fifth embodiment of the present application, a search method is provided. The method includes:

[0026] Responding to a product search event, obtaining product search results;

[0027] Obtaining reference information output by a preset model based on the product search results to assist the user in product selection;

[0028] On the interface, display the search results and the reference information.

[0029] In the sixth embodiment of the present application, a search system is provided. The search system includes:

[0030] A client for executing the steps in the search method provided in the first, second, or third embodiment described above;

[0031] A server for executing the steps in the search method provided in the fourth embodiment described above.

[0032] In the seventh embodiment of the present application, an electronic device is provided. The electronic device includes a memory and a processor. Among them, the memory is used to store executable instructions; the processor runs the executable instructions to implement the steps in the above method embodiments.

[0033] In the eighth embodiment of the present application, a computer-readable storage medium is provided. The storage medium stores computer instructions, and when the instructions are executed by a processor, the steps in the above method embodiments can be implemented.

[0034] The ninth embodiment of the present application further provides a computer program product, which includes a computer program or instructions. When the computer program or instructions are executed by a processor, the processor executes the steps in the above method embodiments.

[0035] In the technical solutions provided by some embodiments of the present application, by utilizing the powerful capabilities of a preset model, the search results are processed to output reference information to assist the user. The reference information can be navigation information sorted out and summarized from the search results, or recommendation information selected from the search results and recommended to the user; the search results obtained based on keyword search and the reference information output by the preset model are displayed together on the interface. In this way, the user can see both the search results and the reference information on the same interface, which is very convenient for the user to understand the situation of the search results, such as what categories are included in the search results, or which result items in the search results are worthy of attention, etc. It can be seen that the technical solutions provided by the embodiments of the present application help to expand the depth and width of the user's search, help the user better understand the search results, reduce the reading cost, and improve the decision-making efficiency.

[0036] In the technical solutions provided by some other embodiments of the present application, a combination of a search engine and a first preset model is used. The search engine is used to search for the first search results related to the keywords; the first preset model is used to identify the user's search intention, and then the search engine is used to search for the second search results related to the search intention; the second preset model is used to process the second search results and output reference information. The reference information can be navigation information sorted out and summarized from the second search results, or recommendation information selected from the second search results and recommended to the user. It can be seen that these embodiments of the present application make full use of the advantages of the search engine and the preset model, retain the powerful functions of the search engine, and also utilize the intention recognition ability of the first preset model to expand the depth and width of the user's search; utilize the sorting and summarizing ability of the second preset model to process the second search results and generate reference information, broadening the search width for the user. The first search results and the reference information corresponding to the second search results provided by artificial intelligence to the user are simultaneously displayed on the interface, providing more information for the user and making the interface content richer. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The schematic embodiments and descriptions thereof of the present application are used to explain the present application and do not constitute an improper limitation to the present application. In the drawings:

[0038] Figure 1 is a schematic structural diagram of a search system provided by an exemplary embodiment of the present application;

[0039] Figure 2 is a schematic flowchart of a search method provided by an exemplary embodiment of the present application;

[0040] Figure 3a 、 3b and 3c are schematic diagrams of multiple interfaces provided by an exemplary embodiment of the present application;

[0041] Figure 4a is a schematic diagram of an interface provided by another exemplary embodiment of the present application;

[0042] Figure 4b is a schematic diagram of page switching provided by an exemplary embodiment of the present application;

[0043] Figure 5 is a schematic flowchart of a search method provided by another exemplary embodiment of the present application;

[0044] Figure 6 is a schematic flowchart of a search method provided by another exemplary embodiment of the present application;

[0045] Figure 7Flow chart of the search method provided by another exemplary embodiment of the present application;

[0046] Figure 8 Flow chart of the search method in the e-commerce scenario provided by an exemplary embodiment of the present application;

[0047] Figure 9 Flow chart of the search method in the e-commerce scenario provided by another exemplary embodiment of the present application;

[0048] Figure 10 Schematic structural diagram of an electronic device provided by an exemplary embodiment of the present application. Detailed implementation manners

[0049] To make the objectives, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be clearly and completely described below in conjunction with specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts shall fall within the scope of protection of the present application.

[0050] It should be noted that in the case where the embodiments of the present application involve user information, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the embodiments of the present application are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions, and corresponding operation entrances are provided for users to choose to authorize or refuse. In addition, various models involved in the present application (including but not limited to language models or large models) comply with relevant laws and standards.

[0051] In addition, it should be noted that in the case where the embodiments of the present application involve user interaction operations or trigger operations, the user interaction operations or trigger operations involved in the embodiments of the present application include but are not limited to: interaction operations in various ways such as touch operations, gesture operations, voice operations, head movement operations, and eye movement operations; among them, touch operations include but are not limited to: click operations, double-click operations, long-press operations, sliding operations, pinch operations or mouse hover operations, etc. Sliding operations include but are not limited to: linear sliding, curved sliding, etc.

[0052] Furthermore, it should be noted that, in the case where the embodiments of the present application involve jumping between the first interface and the second interface, the jumping methods involved in the embodiments of the present application include but are not limited to: jumping directly from the first interface to the second interface, jumping from the first interface to the task interface first and jumping to the second interface after completing the corresponding task operation on the task interface; completing the corresponding task operation on the task interface includes but is not limited to: when the task interface is implemented as an identity authentication interface, completing identity authentication on the identity authentication interface; when the task interface is implemented as a recharge interface, completing the recharge operation on the recharge interface; and so on.

[0053] Before introducing the specific contents of each embodiment of the present application, a brief explanation of the technical terms mentioned in this article is first given.

[0054] Query: Query means query. Query is a query parameter class, which is a message sent by a search engine or database to search for files, pictures, websites, records or a series of records in the database. Query is a query parameter class that encapsulates query conditions, paging, sorting and other functions.

[0055] Preset model: refers to a machine learning model with a large number of parameters generated by training with self-supervised or unsupervised methods using massive data, which provides excellent distributed feature representation and model generalization capabilities for downstream tasks. The preset model can be called a generalized large model, which can include the first preset model, the second preset model, etc. mentioned below; the structure of each model can be different, and the training data can also be different.

[0056] An e-commerce service provider, or e-commerce platform, or simply e-commerce platform, is a platform that provides online transactions and negotiations for enterprises or individuals. An e-commerce service provider is a place that coordinates and integrates the orderly association and flow of information flow, goods flow, and capital flow, that is, by establishing an e-commerce portal, it provides online shopping services to ordinary consumers, or provides business applications to enterprises, in order to seek a better combination between the Internet and traditional commercial and industrial systems. Enterprises, merchants and individuals can make full use of the shared resources such as network infrastructure and payment platforms provided by e-commerce platforms to carry out their own business activities. In this article, e-commerce service providers can be store bazaar e-commerce platforms, shopping mall e-commerce platforms or social e-commerce platforms. Store bazaar e-commerce platforms and shopping mall e-commerce platforms can be called traditional e-commerce. Social e-commerce is a new derivative model of e-commerce. It uses social networking sites, social apps and other communication channels to assist the purchase and sale of goods through social interaction, user-generated content and other means, and applies social elements such as attention, communication, discussion and interaction to the e-commerce transaction process. Social e-commerce integrates e-commerce and social media.

[0057] Merchants refer to the general term for individuals and various organizations engaged in commercial activities (producing and operating related items) through e-commerce platforms.

[0058] SPU (Standard Product Unit) is a standardized product unit used to describe the characteristic information of a product. It is the smallest unit for aggregating product information and is usually composed of a set of reusable and easily retrievable standardized information. The key attributes of SPU determine its uniqueness, and these attributes include brand, model, etc., which can accurately locate a product.

[0059] Navigation information, in this article, refers to the tool that helps users browse and find the required information (such as the result items in search results) on the website. The manifestation forms of navigation can be, but are not limited to: navigation bars, drop-down menus, tabbed layout labels for navigation, and fixed navigation bars (fixed on a certain side of the screen as the page scrolls, facilitating users to access at any time).

[0060] With the rapid development of large model technology, its application prospects in various fields are becoming increasingly broad. It is not common in the field of assisting users to understand search results. Most search websites also use artificial intelligence technology to answer the questions searched by users, rather than organizing the search results to enable users to quickly understand the situation of the search results recalled by the search engine. In the e-commerce field, the use of large models (i.e., artificial intelligence) is mostly in the field of intelligent customer service.

[0061] The inventors of this application innovatively proposed to utilize artificial intelligence technology in the processing of search results. For example, automatically generating navigation information for search results to facilitate users to better understand and recognize search results. In addition, the navigation links generated by using artificial intelligence technology can basically cover most Queries, effectively reducing labor costs (in the prior art, for example, the navigation information on the e-commerce navigation page is manually configured).

[0062] Therefore, this application is proposed. The technical solutions of the embodiments of this application combine artificial intelligence technology with the search engine. The first preset model, the second preset model, etc. mentioned in this article are obtained through training with a large amount of training data, and the training process of the models is not specifically limited in the embodiments of this application.

[0063] In specific implementation, corresponding functional interfaces can be deployed on the front-end application (APP) or browser web page so that after the user triggers a corresponding operation (such as a search operation), the corresponding services (such as the search service provided by the search engine, the processing service of the preset model, etc.) of the back-end (such as the server) can be automatically called through the functional interface. The form of the functional interface on the APP page or browser web page can be a functional control, and the user can call the corresponding services of the back-end by clicking the functional control.

[0064] The technical solutions provided by the embodiments of the present application need to be implemented under the system architecture as shown in Figure 1 This embodiment provides a search system, which includes a client 2 and a server 1. Among them, the server 1 can be a server, a service cluster, a virtual server, or the cloud, etc., and this embodiment does not make specific limitations on this. The server 1 provides corresponding functional services for the client 2, such as search services, artificial intelligence services, etc. At least one preset model is deployed on the server 1. For example, a preset model for identifying the user's search intention is deployed, a preset model for processing the search results to obtain reference information is deployed, etc. Of course, the server 1 can also only deploy a powerful preset model, which can not only identify the user's search intention, but also automatically construct navigation information for the search results, and can also intelligently recommend content to the user, etc. The user can enter the search interface through a browser, a client application (APP), a web application H5 (HyperText Markup Language 5, the fifth generation HTML, hypertext markup language), a light application (also known as a small program, a lightweight application program), or a cloud application on the corresponding device of the client 2. The client device can be, but is not limited to: a smart phone, a smart wearable device, a tablet computer, a laptop computer, a desktop computer, etc. In this embodiment, the client can implement the steps in the methods provided by some of the following method embodiments, and the server can implement the steps in the methods provided by some of the following method embodiments, which will be specifically described below.

[0065] Figure 2 shows a schematic flow chart of a search method provided by an embodiment of the present application. As shown in Figure 2 This embodiment provides a method whose execution subject can be the client 2. The method includes:

[0066] 101. Respond to a search event and obtain search results related to the keyword;

[0067] 102. Obtain reference information for assisting the user to understand the search results output by the preset model based on the search results;

[0068] 103. Display the search results and the reference information on the interface.

[0069] In the above 101, the search results may include multiple result items matched by the search engine based on the keyword, and / or use artificial intelligence technology to analyze the keyword to obtain the user's search intention, and generate at least one search term based on the search intention, and the search engine searches based on at least one search term to obtain multiple result items.

[0070] In the above 102, the reference information can be navigation information corresponding to the search results, or recommendation information based on the search results.

[0071] For example, asFigure 3a As shown, the preset model generates navigation information 3 by analyzing and classifying search results. The navigation information includes grouping titles 4 of multiple classification groups, and the result items included in the classification groups correspond to page elements 5. As Figure 3b shown in the example, the preset model selects at least one recommended result item from the search results and generates recommendation information 6 based on the at least one recommended result item.

[0072] In the above 103, the search results and reference information can be displayed in two different areas on the interface; the two areas can be adjacent. For example, the search results are displayed in the second area and the reference information is displayed in the first area. The second area can be located below the first area. Or, a search box 7 is also displayed on the interface. As Figure 3a and 3b shown, the first area is located between the search box and the second area. When the user operates in the second area and the content of the second area scrolls, the position of the second area can remain unchanged.

[0073] In the technical solution provided in this embodiment, by using the powerful ability of the preset model, the search results are processed to output reference information to assist the user. The reference information can be navigation information obtained by sorting and summarizing the search results, or recommendation information selected from the search results and recommended to the user. The search results obtained based on keyword search and the reference information output by the preset model are displayed on the interface together. In this way, the user can see both the search results and the reference information on the same interface, which is very convenient for the user to understand the situation of the search results, such as what categories are in the search results, or which result items in the search results are worthy of attention, etc. It can be seen that the technical solution provided in the embodiment of the present application helps to expand the depth and width of the user's search, helps the user better understand the search results, reduces the reading cost, and improves the decision-making efficiency.

[0074] In a specific implementation example, the reference information can be navigation information. As Figure 3a shown, the navigation information includes: multiple classification titles obtained by the preset model analyzing and classifying multiple result items in the search results; and correspondingly, "displaying the navigation information" in the above step 103 can include:

[0075] 1031. Display first operable controls respectively corresponding to the multiple classification titles.

[0076] Further, the navigation information further includes: at least one result item in the classification group corresponding to the classification title; and the method further includes:

[0077] 104. In response to the triggering of a first operable control corresponding to a target classification title among the multiple classification titles, display page elements of at least some result items in the classification group corresponding to the target classification title. Among them, the page elements of the result items include at least one of a picture of the result item, a video, a difference range in which the same result item provided by different providers is different as summarized by the preset model, and characteristic information of the result item as summarized by the preset model.

[0078] Alternatively, the navigation information includes multiple sub-group sets included in the classification group, the sub-group set includes at least one result item, and the method further includes:

[0079] 104'. In response to the triggering of a first operable control corresponding to a target classification title among the multiple classification titles, display page elements of at least some sub-group sets in the classification group corresponding to the target classification title;

[0080] Among them, the page elements of the sub-group set include at least one of a picture representing the result items in the sub-group set, a video, a difference range in which the result items included in the sub-group set are different as summarized by the preset model, and characteristic information of the result items included in the sub-group set as summarized by the preset model.

[0081] In specific implementation, the multiple result items included in the classification group can be regrouped to divide out multiple sub-group sets. For example, in an e-commerce scenario, the result items can be products, and the products have SPU. The result items included in the classification group can be grouped according to the SPU, that is, the products with the same SPU are aggregated into a sub-group set.

[0082] See Figure 3a In the shown example, the page elements of the result items include at least one of the theme 85 of the result item, the picture 81, the video 82, the difference range 83 in which the same result item provided by different providers is different as summarized by the preset model, and the characteristic information 84 of the result item as summarized by the preset model. Or, Figure 3a In the shown example, the page elements of the sub-group set include at least one of a picture representing the result items in the sub-group set, a video, a difference range in which the result items included in the sub-group set are different as summarized by the preset model, and characteristic information of the result items included in the sub-group set as summarized by the preset model.

[0083] The user operates on the interface as Figure 3a shown. The user clicks the first operable control corresponding to the target classification title "Brand A", and at this time, page elements of at least some result items in the classification group corresponding to "Brand A" are displayed. Before the user operates, as Figure 3aAs shown, the currently displayed is the first operable control corresponding to "Super Bestseller", and the first operable control corresponding to "Super Bestseller" is highlighted, such as in bold, large font size, etc. After the user operates and clicks the first operable control corresponding to the target classification title "Brand A", as Figure 3c shown, the first operable control corresponding to "Super Bestseller" resumes to be displayed in the set font, set font size, etc.; while the first operable control corresponding to "Brand A" is highlighted, such as in bold, large font size, etc. At the same time, the page elements displayed below multiple grouping titles (such as Super Bestseller, Brand A, Brand B, and Brand C) are switched to display the page elements of at least some result items in the classification group corresponding to "Brand A".

[0084] Furthermore, as Figure 3a and 3c shown, a third operable control is also displayed on the interface, such as the control shown in the figure. The user can operate this third operable control to bring out the page elements of the hidden result items or sub-group sets. The third operable control can be set beside the last page element among the page elements of at least some result items or sub-group sets. Suppose the page elements of at least some result items or sub-group sets are arranged horizontally as Figure 3a and 3c shown, then the third operable control can be set on the right side of the last page element among the page elements of at least some result items or sub-group sets. If the page elements of at least some result items or sub-group sets are arranged vertically, then the third operable control can be set below the last page element among the page elements of at least some result items or sub-group sets.

[0085] The difference items for different types of result items are different. For example, for products, the difference item may be the price, and the corresponding difference range may be the price range of the result items within the classification group classified as a group. Still for products, more specifically, such as down jackets, the difference item may be the filling amount, and the corresponding difference range may be the filling amount range of the result items within the classification group classified as a group.

[0086] The characteristic information 84 of the result item can be generated by a preset model based on the relevant information of the result item. The sources of the relevant information of the result item can include the information stored in the server database, and can also include the information searched from the network side, etc. Taking the e-commerce scenario as an example, the object of search is products, and the result items included in the classification group are products; correspondingly, the information of this product can be the information stored on the e-commerce platform, such as the transaction volume of this product, the evaluation information of the product, the product description information, etc.; it can also include the evaluation information searched from the network side about this product, etc. This embodiment does not make specific limitations on this.

[0087] For the result items of non-standard products, such as non-standard product items (i.e., products without SPU information), a preset model can be used to analyze the search results related to the keyword retrieved by the search engine to determine the search intent (such as a general intent); and at least one search term can be determined based on the search intent. The search engine retrieves search results based on the at least one search term; at this time, the search results can be grouped according to the search terms, that is, the search result items corresponding to one search term are grouped into one group, and at least one classification group is obtained.

[0088] The preset model can also summarize the result items within each classification group and then group them to obtain multiple sub-group sets within the classification group. In a specific implementation, the preset model can choose to classify the result items within the classification group in the SPU classification manner, or classify them by analyzing the relevant information of the result items within the classification group.

[0089] Taking the result item as a product as an example, the preset model can classify by analyzing the title of the product, product description information, etc. For example, products with the same or similar target words in the product title can be grouped into one group, or products with the same or similar function descriptions included in the product description information can be grouped into one group; etc., and this embodiment does not make specific limitations on this.

[0090] Further, "displaying the navigation information" in step 103 above may further include:

[0091] Displaying page elements of at least some of the result items in the classification group corresponding to at least one of the multiple classification titles; wherein, the page elements of the result items include at least one of a picture of the result item, a video, the attribute interval summarized by the preset model, and the characteristic information of the result item summarized by the preset model.

[0092] Alternatively, "displaying the navigation information" in step 103 above may further include:

[0093] Displaying page elements of at least some of the sub-group sets in the classification group corresponding to at least one of the multiple classification titles; wherein, the sub-group set includes at least one result item, and the page elements of the sub-group set include at least one of a picture representing the result items in the sub-group set, a video, the difference interval where the result items included in the sub-group set differ, and the characteristic information of the result items included in the sub-group set summarized by the preset model.

[0094] Such as Figure 3a and 3cAs shown, on the interface, there is a page element that shows at least some result items or sub-group sets in a classification group corresponding to a classification title. In fact, if there are not many classification groups and the space they occupy on the page is not much, it is also possible to show more page elements of at least some result items or sub-group sets in the classification groups corresponding to several classification titles. For example, as shown in 3d, on the interface, such as in the first area of the interface, there are page elements of at least some result items or sub-group sets in two classification groups corresponding to two classification titles respectively.

[0095] In specific implementation, the page element of the result item is operable. Correspondingly, the method provided in this embodiment further includes:

[0096] 105. In response to an operation on a page element of a result item or a sub-group set on the interface, display a drop-down menu, and the drop-down menu shows second operable parts of the result items in the result item or sub-group set provided by different providers; wherein, operating the second operable part can trigger the page to jump to the first page corresponding to the result item provided by the corresponding provider; or

[0097] 105’. In response to an operation on a page element of a result item or a sub-group set on the interface, jump to a second page that shows the result items in the result item or sub-group set provided by different providers.

[0098] Regarding the above step 105, refer to Figure 4a As shown, after the user clicks on the page element of a result item, a drop-down menu 9 is shown below the page element, and the drop-down menu shows second operation controls of the result item provided by different providers. The user can trigger the jump to the first page corresponding to the result item provided by the provider by operating the second operation control. Taking an e-commerce scenario as an example, the provider of the result item can be understood as a merchant, and the result item is the product sold by the merchant on the e-commerce platform. There may be multiple merchants on the e-commerce platform selling this product, and the second operation controls of this product provided by different merchants can be shown in the drop-down menu 9. After the user operates the second operation control, they can enter the detail page of this product sold in the merchant's store. The detail page can show product pictures, videos, prices, names, origins, product parameters, user evaluations, etc.

[0099] The display space of the drop-down menu 9 is limited, and the user can also, through operations such as swiping, bring out the hidden second operation controls of the result items. The drop-down menu 9 can show simplified page elements of the result item provided by each provider, such as only including the picture or video of the product, as well as the price, etc. This embodiment does not make specific limitations on this.

[0100] Or, as in the above step 105’, as Figure 4bAs shown, after the user clicks on a page element of a result item, it jumps to the second page 10 that shows the result items provided by different providers. As Figure 4b In the example shown, on the second page 10, page elements of products provided by different merchants (stores) are shown (such as the cards shown in the figure). When the user clicks on a page element of a certain product (such as the card shown in the figure), they can jump to the detail page of the products provided by that store.

[0101] Furthermore, a fourth operable control may also be shown on the second page 10. Such as Figure 4b In the example shown, a "return" control. By touching this "return" control, the user can return to Figure 4b the upper interface in

[0102] In another specific implementation example, the reference information is recommendation information, and the recommendation information includes at least one recommended result item selected by the preset model from the search results and the recommended text generated for the recommended result item. Correspondingly, "showing the recommendation information" in the above 103 may include:

[0103] Showing page elements of at least some of the recommended result items in the recommendation information;

[0104] Among them, the page elements of the recommended result item include: the title of the recommended result item, the recommended text, and the picture or video of the recommended result item.

[0105] Such as Figure 3b In the example shown, page elements of at least some of the recommended result items in the recommendation information are shown on the interface. Similarly, the page elements of the recommended result item can also be operable. By clicking on the page element of a certain recommended result item, the user can call up the same Figure 4a drop-down menu 9 shown, or jump to the second page that shows the recommended result items provided by different providers as shown in 4b. Similarly, a fourth operable control may also be provided on the second page, similar to Figure 4b the "return" control shown in Figure 3b the interface shown, and can return from the second page

[0106] Continuing to refer to Figure 3b In the example shown, the page elements of the recommended result item include: the title 113 of the recommended result item, the recommended text 114, and the picture 111 or video 112 of the recommended result item.

[0107] Furthermore, the method provided in this embodiment may further include the following steps:

[0108] Obtaining the difference interval where different providers providing the recommended result item are different based on the relevant information of the recommended result item by the preset model;

[0109] Correspondingly, the page element of the displayed recommended result item further includes a difference range 115 (as shown in the example Figure 3b illustrated). In an e-commerce scenario, the difference range can be the price range of a product, the parameter range of a product (such as the down filling amount range of a down jacket, the memory capacity range of a smartphone, etc.).

[0110] In specific implementation, the recommended text of the recommended result item can be generated by a preset model based on the relevant information of the recommended result item. The relevant information of the recommended result item may include, but is not limited to: product parameters, product evaluation information, product transaction volume, product search volume, product description information, product advertisements, etc. The relevant information of the recommended result item can come from the data stored on the server side or the information searched from the network side. This embodiment does not make specific limitations on this.

[0111] Figure 5 FIG. shows a schematic flowchart of a search method provided by another embodiment of the present application. As Figure 5 shown, the execution subject of the method provided in this embodiment can be a client. Specifically, the method may include:

[0112] 201. Respond to a search event and obtain search results related to a keyword, where the search results are obtained by searching based on a search intention related to the keyword, and the search intention is generated by a first preset model based on the keyword;

[0113] 202. Obtain reference information output by a second preset model based on the search results;

[0114] 203. Display the search results and the reference information on the interface;

[0115] wherein, the first preset model and the second preset model are the same model or two different models; the reference information is navigation information for assisting the user to understand the search results or recommendation information; the recommendation information includes at least one recommended result item selected by the second preset model from the search results and the recommended text generated for the recommended result item.

[0116] In the technical solution provided in this embodiment, by leveraging the powerful capabilities of the first preset model, the search intent is determined based on keywords, and then search results are obtained based on the search intent. Compared with the search method using keyword matching, the search results are more comprehensive and accurate, which is beneficial to broadening the search breadth. Further, the solution provided in this embodiment uses the second preset model to process the search results and output reference information to assist the user. The reference information can be navigation information sorted out and summarized from the search results, or recommended information selected from the search results and recommended to the user. The search results obtained based on keywords and the reference information output by the preset model are displayed on the interface together. In this way, the user can see both the search results and the reference information on the same interface, which is very convenient for the user to understand the situation of the search results, such as what categories are included in the search results, or which result items in the search results are worthy of attention, etc. It can be seen that the technical solution provided in the embodiment of this application helps to expand the depth and breadth of the user's search, helps the user better understand the search results, reduces the reading cost, and improves the decision-making efficiency.

[0117] It should be noted here that: for a more detailed description of the above steps 202 and 203, please refer to the above text. In addition, this embodiment may also include the steps in the above text. Some steps in the above embodiment can be directly introduced into this embodiment, and will not be elaborated here.

[0118] Figure 6 The flowchart of the search method provided in another embodiment of this application is shown. As Figure 6 shown, the execution subject of the method provided in this embodiment can be a client. Specifically, the method may include:

[0119] 301. Respond to a search event and obtain a first search result related to the keyword retrieved by a search engine.

[0120] 302. Obtain a second search result related to the keyword, where the second search result is retrieved by the search engine based on a search intent related to the keyword, and the search intent is generated by a first preset model based on the keyword.

[0121] 303. Obtain first reference information output by a second preset model based on the second search result.

[0122] 304. On the interface, display the first search result and the first reference information.

[0123] Among them, the second preset model may be trained. The first preset model and the second preset model are the same model or two different models. The first reference information is the first navigation information or the first recommendation information for assisting the user to understand the second search result. The first recommendation information includes at least one recommended result item selected by the second preset model from the second search result and the recommended text generated for the recommended result item.

[0124] In the technical solution provided in this embodiment, by combining a search engine and a first preset model, the search engine is used to search for the first search result related to the keyword. The first preset model is used to identify the user's search intention, and then the search engine is used to search for the second search result related to the search intention. Then, the second preset model is used to process the second search result and output reference information. The reference information can be the navigation information sorted out and summarized from the second search result, or the recommendation information recommended to the user selected from the second search result. It can be seen that these embodiments of the present application make full use of the advantages of the search engine and the preset model, retain the powerful function of the search engine, and also utilize the intention recognition ability of the first preset model to expand the depth and width of the user's search. By using the sorting and summarization ability of the second preset model, the second search result is processed to generate reference information, which broadens the search width for the user. The first search result and the reference information corresponding to the second search result provided by the artificial intelligence are simultaneously displayed on the interface, providing more information for the user and making the interface content richer.

[0125] Further, the above step 303: "Obtain the first reference information output by the second preset model based on the second search result" may include:

[0126] Obtain the second reference information output by the second preset model based on the first search result and the second search result. Among them, the second reference information is the second navigation information for assisting the user to understand the first search result and the second search result, or the second recommendation information. The second recommendation information includes at least one recommended result item selected by the preset model from the first search result and the second search result and the recommended text generated for the recommended result item. Correspondingly, the above step 304 "Display the first search result and the first reference information on the interface" may be specifically:

[0127] Display the first search result, the second search result, and the second reference information on the interface.

[0128] It should be noted here that for a more detailed description of the above steps 303 and 304, please refer to the above text. In addition, this embodiment may also include the steps in the above text. Some steps in the above embodiment can be directly introduced into this embodiment and will not be elaborated here.

[0129] The methods provided in the above embodiments can be applied to the e-commerce scenario. That is, another embodiment of the present application provides a product search method, including:

[0130] 401. Respond to the product search event and obtain the product search results;

[0131] 402. Obtain the reference information for assisting the user in product selection output by the preset model based on the product search results;

[0132] 403. On the interface, display the search results and the reference information.

[0133] It should be noted here that for a more detailed description of the above steps 401 to 403, reference can be made to the above text. In addition, this embodiment may also include the steps in the above text. Some steps in the above embodiments can be directly introduced into this embodiment, and the result items in the above embodiments can be replaced with products, which will not be elaborated here.

[0134] Figure 7 The flowchart shows a search method provided by another embodiment of the present application. The execution subject of the method provided in this embodiment can be the server. Specifically, the method may include:

[0135] 501. Obtain the keyword;

[0136] 502. Search for the search results related to the keyword;

[0137] 503. Use the preset model to process the search results to obtain the reference information for assisting the user in understanding the search results;

[0138] 504. Send the search results and the reference information to the client for display on the client interface.

[0139] In the above 501, the keyword can be the input information input by the user through the client, sent to the server by the client, and then determined by the server based on the input information. It should be supplemented here that the "keyword" mentioned in each embodiment in this article may include at least one term. That is, the server processes the user input information to obtain a keyword containing at least one term.

[0140] Furthermore, the method provided in this embodiment may further include the following steps:

[0141] 505. Use the preset model to analyze the keyword to determine the search intention;

[0142] 506. Generate at least one search term based on the search intention.

[0143] Further, in the method embodiment where the execution entity is a client as described above, the client may also send the user's identification information to the server. The server may obtain the user's historical behavior data based on the user's identification information, etc., and use the user's historical behavior data and the keyword as the input of the preset model to execute the preset model to analyze the user's search intent.

[0144] Correspondingly, step 502 "search for search results related to the keyword" may be specifically:

[0145] Use a search engine to search for a first result set related to the keyword; and / or

[0146] Use a search engine to search for a second result set related to the at least one search term;

[0147] Wherein, the search results include the first result set and / or the second result set.

[0148] Alternatively, step 502 "search for search results related to the keyword" may include the following steps:

[0149] 5021. Use a search engine to search for a first result set related to the keyword;

[0150] 5022. Use a preset model to determine the search intent based on the first result set;

[0151] 5023. Generate at least one search term based on the search intent;

[0152] 5024. Use a search engine to search for a second result set related to the at least one search term;

[0153] 5025. Determine the search results according to the first result set and the second result set.

[0154] In a specific implementation example, the reference information may be navigation information. Correspondingly, step 503 "use a preset model to process the search results to obtain navigation information" includes:

[0155] 5031. The preset model analyzes the search results to determine the classification method;

[0156] 5032. The preset model classifies the search results according to the classification method to obtain multiple classification groups and corresponding classification titles.

[0157] Among them, the classification method can also be understood as an aggregation dimension. Taking the e-commerce scenario as an example, the search results include product information of multiple products. The classification method can be to classify according to the product brand, that is, group the products with the same brand in the search results into a group as a classification group, and then multiple classification groups corresponding to different brands can be obtained. Correspondingly, the classification title corresponding to the classification group can be the brand name. Or, the classification method can be to classify according to the product category. Suppose

[0158] Further, the above step 503, "using a preset model to process the search results to obtain navigation information", further includes:

[0159] 5033. The preset model analyzes the result items included in the classification group, and summarizes the characteristic information of the result items and / or the difference range where the same result items provided by different providers are different.

[0160] For example, when the result items in the search results are products, the classification method includes at least one of the following: product brand, product category, and the standardized product unit SPU of the product.

[0161] Suppose there are multiple alternative classification methods. Correspondingly, the above step 5021, "the preset model analyzes the search results to determine the classification method", may include:

[0162] The preset model analyzes the search results and selects one from the multiple alternative classification methods based on the analysis results; or

[0163] The preset model selects one from the multiple alternative classification methods according to the search intent and the search results; wherein, the search intent is determined by the preset model based on the keywords.

[0164] Further, the method provided in this embodiment may further include the following steps:

[0165] Perform a quality assessment on the multiple classification groups obtained according to the selected classification method;

[0166] When the quality assessment passes, trigger the step that the preset model analyzes the result items included in the classification group and summarizes the characteristic information of the result items and / or the difference range where the same result items provided by different providers are different; or group the result items included in the classification group to obtain a subset of sub-classification groups within the classification group, sort the multiple subsets of sub-classification groups within the group, and summarize the characteristic information of the result items within the subset of sub-classification groups and / or the difference range where the result items within the subset of sub-classification groups are different;

[0167] When the quality assessment fails, the preset model selects another one from the multiple alternative classification methods.

[0168] Further, when multiple classification groups obtained by classifying the search results according to the re-selected classification method still fail to pass the quality assessment, the method provided in this embodiment further includes:

[0169] Using the search terms as the classification method, aggregating the search results corresponding to different search terms into one classification group; wherein, the multiple search terms are determined by a preset model through intent recognition of keywords.

[0170] That is, in the solution provided in this embodiment, when multiple classification groups obtained by classifying the search results according to the re-selected classification method still cannot pass the quality review, aggregation is performed according to the search terms. For example, referring to the preset model mentioned above to analyze the keywords to determine the search intent, and generating multiple search terms based on the search intent, the search engine searches for the search results corresponding to each search term according to the multiple search terms respectively. At this time, the search results corresponding to the search terms are used as one classification group. For example, if the search results corresponding to the first search term are {A1, A2, A3,...}, then {A1, A2, A3,...} is used as one classification group. If the search results corresponding to the second search term are {B1, B2, B3,...}, then {B1, B2, B3,...} is used as another classification group.

[0171] In another specific implementation example, the reference information may be recommendation information. Correspondingly, the step "using a preset model to process the search results to obtain recommendation information" in this embodiment may include:

[0172] The preset model analyzes the search results, selects at least one recommended result item from the search results, and generates a recommended text for the recommended result item.

[0173] For the convenience of understanding, the present application also provides an embodiment taking an e-commerce scenario as an example. The search method provided in this embodiment is based on the intent recognition of a preset model, uses intent as a supplement, and combines the information aggregation of the search engine. On the one hand, it can quickly and accurately generate navigation information for users in a single search in real time, provide users with more input of product knowledge, and help users better understand and recognize products. On the other hand, this automated link generation solution provided by this solution can basically cover most Queries, effectively releasing the labor cost of operations. Generally speaking, the solution provided in this embodiment features "I (AI) help you (user) read the first N products on the search result page and tell you that there are probably these types", so as to expand the search depth and width of users, help users better understand products, reduce the reading cost, and improve the decision-making efficiency.

[0174] In the traditional e-commerce scenario, when users have shopping needs, they will come to the e-commerce platform and retrieve corresponding products by entering keywords. However, in some scenarios with relatively broad needs (where users' needs are category words such as headphones or scenario words such as camping), the products searched out will be very diverse, spanning categories / brands / models, resulting in users simply "not knowing how to choose". At this time, users generally click on multiple products to view the details and parameters, and at the same time search for relevant introductions outside the platform to understand the products. Therefore, for users, the entire shopping path is cumbersome and complex. When shopping, users expect someone to tell them about the market situation, such as which models have high sales, which models have high positive reviews, what the price ranges are distributed like, and what the characteristics of each model are... and so on.

[0175] Based on such a background, for the decision-making problems encountered by users during the selection process, in this embodiment, through the method of preset model + search engine aggregation, it mainly features "I (AI) help you (the user) read the first N products on the search result page and tell you that there are approximately these types", so as to expand the search depth and width of users, help users better understand products, reduce the reading cost, and improve the decision-making efficiency.

[0176] Traditional shopping guide products - graphic navigation and the solution of this embodiment aim to solve the same type of user problems. However, traditional graphic navigation is basically configured manually, which leads to relatively high requirements for personnel capabilities and high labor costs on the one hand; on the other hand, it can only cover some head queries, and some medium and long-tail queries are missing. And the combined search demand for medium and long-tail queries in daily life is very large, which is a part that urgently needs to be solved. Manually configured navigation requires the operation team to have profound knowledge and meticulous operation capabilities. Over time, the navigation content needs to be continuously updated and maintained, which increases the labor cost and time cost. Among them, long-tail words are those with a small daily search volume but with continuous search volume.

[0177] The search method provided in this embodiment can at least partially solve the above problems. Specifically, as Figure 8 shown, the search method may include:

[0178] S1. The client responds to the input operation of the user on the interface and determines the keyword.

[0179] Of course, in specific implementation, the client can also send the input information entered by the user to the server, and the server determines the keyword based on the input information.

[0180] S2. The server obtains the keyword, and the search engine recalls the product set based on the keyword to obtain the search result.

[0181] Of course, in step S2, a preset model can also be used to identify the user's search intent. For example, the preset model identifies the user's search intent based on the user's input information; or the client sends the user identifier and the user's input information to the server together, and the server can obtain the user behavior data that has been permitted to be used by the user based on the user identifier, etc.; the user behavior data and the user's input information are used as the input of the preset model, and the preset model is executed to obtain the search intent.

[0182] Then, the preset model generates at least one search term based on the search intent. The search engine recalls a set of products based on the at least one search term. These sets of products recalled based on the at least one search term can also be added to the search results.

[0183] S3. Determine whether the result item in the search results is a standard product. If so, select one classification method from multiple alternative classification methods as the target classification method, and execute steps S4 to S8; otherwise, execute step S9.

[0184] Among them, a standard product refers to a product with SPU information.

[0185] Therefore, the preset model can select a classification method based on the characteristics of the products in the search results. For example, the product is a standard product with SPU information. The preset model can select a classification method classified by brand, or classified by SPU, or classified by category, or classified by product sales volume (that is, products with the same sales volume range are grouped together), etc. For example, Figure 3a The "super bestseller" in

[0186] For non-standard products, such as products without SPU information, the preset model + RAG method can be used to let the preset model output corresponding hyponyms according to the user's search intent. A hyponym refers to a word with a narrower connotation and a smaller extension conceptually, and the things it represents can be completely included by another word (i.e., the hypernym). The hyponyms generated by the preset model aggregate different products together.

[0187] For example, the user's input information is "summer women's dresses"; based on the products included in the search results (such as long skirts, short skirts, skirt pants, long dresses, short dresses, jumpsuits, etc.), the preset model can determine the corresponding hyponyms as: dresses, skirts; in this way, the preset model can select the determined hyponyms as the classification method, such as dividing the dresses in the search results into a classification group with the corresponding classification title as dresses, and dividing the skirts in the search results into a classification group with the corresponding classification title as skirts.

[0188] S4. Classify the products in the search results according to the target classification method to obtain multiple classification groups and the classification titles of each classification group.

[0189] If the selected classification method is the brand classification method, then according to the brand classification method, group the products of the same brand in the search results into one group to obtain multiple brand classification groups and the classification titles of each classification group (which can be brand names).

[0190] If the selected classification method is the category classification method, then according to the category classification method, group the products of the same category in the search results into one group to obtain multiple category classification groups and the classification titles of each classification group (which can be category names). For example, the clothing category may include but is not limited to: tops, pants, skirts, etc.

[0191] If the selected classification method is the SPU classification method, then according to the SPU classification method, group the products with the same SPU in the search results into one group to obtain multiple SPU classification groups and the classification titles of each classification group (which can be model names).

[0192] S5. Conduct quality assessment on multiple classification groups.

[0193] In specific implementation, some evaluation rules can be preset in advance, and when conducting quality assessment on multiple classification groups, the quality assessment can be carried out based on the preset evaluation rules. The specific implementation of the quality assessment in this embodiment is not specifically limited.

[0194] When the assessment passes, the products within each classification group are grouped again to obtain the sub-group sets within each classification group; and sort the multiple sub-group sets within the classification group so that the top N items in the sorting can be displayed on the client side interface.

[0195] For example, in step S4 above, the brand A classification group, brand B classification group, and brand C classification group are separated using the brand classification method, and the quality assessment of each brand classification group passes. In this step S6, the products within each brand classification group are grouped again. Suppose the brand A classification group includes (product C1, product C2, product C3, product C4, product C5). In this step, product C1, product C2, and product C3 can be grouped into one group to obtain the first sub-group set, and product C4 and product C5 can be grouped into one group to obtain the second sub-group set. When grouping the sub-group sets, the SPU grouping method can be used. For example, products of the same model are grouped into one sub-group set.

[0196] See Figure 3cIn the example shown, classification titles of multiple classification groups are displayed on the interface, such as Brand A, Brand B, and Brand C. Below the multiple classification titles, multiple sub-group sets included in the Brand A classification group are displayed, and each sub-group set is displayed on the interface in a card manner. For example, the Brand A classification group includes multiple sub-group sets, and each sub-group set contains products of the same model. The solution provided in this embodiment also sorts the multiple sub-group sets so that the card-style page elements corresponding to the sub-group sets sorted earlier can be arranged in the front. The sub-group set of Model E is sorted first, the sub-group set of Model F is sorted second, the sub-group set of Model G is sorted third, the sub-group set of Model H is sorted fourth,.... As Figure 3c In the example shown, from left to right, the card-style page elements corresponding to the sub-group set of Model F, the card-style page elements corresponding to the sub-group set of Model F, the card-style page elements corresponding to the sub-group set of Model G, and the card-style page elements corresponding to the sub-group set of Model H are arranged in sequence.

[0197] The multiple products included in the sub-group set can be sorted or not sorted. Refer to Figure 4a As shown, after the user clicks on the card-style page element corresponding to a certain sub-group set, the page elements corresponding to the products included in the sub-group set are displayed in the drop-down menu. Or, as Figure 4b shown, after the user clicks on the card-style page element corresponding to a certain sub-group set, it jumps to the second page, and the page elements corresponding to the products in the sub-group set are displayed on the second page.

[0198] S7. A preset model generates characteristic information, price ranges, etc. for each sub-group set.

[0199] Among them, the characteristic information may include but is not limited to: high cost performance, bestseller, top N in the hot search list, high-end model, suitable for home or commercial use, etc. Briefly understood, the characteristic information can be understood as the selling points of the products in the sub-group set.

[0200] The products in the sub-group set can include those provided by different merchants on the e-commerce platform, and may also include products of different sub-models under the same SPU.

[0201] Of course, in specific implementation, multiple alternative classification methods can be selected once, and the search results are classified respectively according to the multiple alternative classification methods to obtain multiple classification groups and the classification titles of each classification group corresponding to each classification method.

[0202] Suppose multiple alternative classification methods include a first classification method and a second classification method. Classify the search results according to the first classification method to obtain a first classification result; classify the search results according to the second classification method to obtain a second classification result. The classification results include multiple classification groups and the classification titles of each classification group. If both the first classification result and the second classification result pass the quality assessment, then perform the above steps S6 and S7 respectively for the first classification result.

[0203] S8. When the evaluation fails, abandon pushing the navigation information for assisting the user to the client.

[0204] That is, the server does not push navigation information similar to Figure 3a the navigation information 3 shown in

[0205] S9. The preset model determines the general intention.

[0206] The preset model can determine the general intention based on the user's input information, or based on the user's input information and the user's historical behavior information, or based on the search results related to the keyword searched by the search engine. Among them, the keyword is determined according to the user's input information.

[0207] After the user inputs information, the process of presetting the generation of the general intention involves the understanding of the user input and the recognition of the intention. The general intention usually refers to a relatively broad or general intention, which reflects the overall purpose or expectation that the user may hold when conducting a search. The general intention is usually abstracted from multiple specific intentions and reflects the overall trend or purpose of the user's search behavior.

[0208] In addition to directly based on the user's input information, the preset model may also consider context information such as the user's search history and the content of the current page to further refine the general intention. Context understanding helps the preset model to more accurately grasp the user's true needs and improve the accuracy and relevance of the search results.

[0209] Or, the search engine recalls search results based on the user's input information. The preset model can determine the general intention by analyzing the search results. Generally speaking, the preset model determines the general intention by summarizing information such as the titles of at least some result items (such as the top N structural items) in the search results.

[0210] S10. The preset model generates multiple search terms based on the general intention.

[0211] S11. Use the search engine to search for the search results corresponding to each search term respectively based on the multiple search terms.

[0212] S12. Use the search term as a classification method to aggregate the search results corresponding to one search term into a classification group.

[0213] S13. Group the products within the classification group according to the SPU grouping method to obtain a subset of sub - groups within the classification group, and sort the multiple subsets of sub - groups within the classification group so that the top N items in the sorting can be displayed on the interface, and then return to execute step S7.

[0214] Another embodiment of the present application, taking an e - commerce scenario as an example, provides a search method, which includes:

[0215] S1. The client responds to the user's input operation on the interface to determine a keyword.

[0216] S2. The server obtains the keyword, and the search engine recalls a set of products based on the keyword to obtain search results.

[0217] S3'. When the result item in the search results is a standard product, select one classification method from multiple alternative classification methods as the target classification method.

[0218] S4. Classify the products in the search results according to the target classification method to obtain multiple classification groups and the classification titles of each classification group.

[0219] S5. Conduct a quality assessment on the multiple classification groups.

[0220] S6. When the assessment passes, the products within each classification group are further grouped to obtain a subset of sub - groups within each classification group; and sort the multiple subsets of sub - groups within the classification group so that the top N items in the sorting can be displayed on the interface.

[0221] S7. The preset model generates characteristic information, price ranges, etc. for each subset of sub - groups.

[0222] S8'. When the assessment fails, select another one from the multiple alternative classification methods through a preset selection rule as the target classification method, and execute steps S4 to S7; if at least one recommended result item selected according to the re - selected classification method still fails the assessment, then execute steps S9 - S13 in the above - mentioned embodiment.

[0223] As Figure 9 shown, another embodiment of the present application, taking an e - commerce scenario as an example, provides a search method, which includes:

[0224] S1. The client responds to the user's input operation on the interface to determine a keyword.

[0225] S9. The preset model determines a general intention.

[0226] S10. The preset model generates multiple search terms based on the general intention.

[0227] S11. Use a search engine to search for the search results corresponding to each search term based on the multiple search terms respectively.

[0228] S12. Use the search terms as a classification method to aggregate the search results corresponding to one search term into a classification group.

[0229] S13. Group the products within the classification group according to the SPU grouping method to obtain a subset of sub - groups within the classification group, and sort the multiple subsets of sub - groups within the classification group so that the top N items in the sorting can be displayed on the interface when displayed on the client side.

[0230] S7. A preset model generates characteristic information, price range, etc. for each subset of sub - groups.

[0231] When the preset model generates characteristic information for the subset of sub - groups, it can combine with the RAG technology.

[0232] The technical solution provided in this embodiment can use the combination of a large model + RAG (Retrieval Augmented Generation) to achieve real - time and batch production, and can effectively release the cost of operation and manual configuration. The RAG technology solves the problems of knowledge extraction and recall by building a retrieval system outside the large model. It uses the efficient storage and retrieval capabilities of the vector database to recall knowledge related to the target query and integrates it into the Prompt, enabling the preset model to give reasonable answers with reference to this knowledge. For example, the RAG technology can be used to collect a large amount of text data such as product descriptions, user reviews, and professional evaluation articles related to the SPU; use the search engine technology to index the text data; retrieve several text fragments related to the products within the subset of sub - groups through the search engine; and sort the search results according to factors such as text relevance, timeliness, and authority. Then, use the preset model to understand and integrate the retrieved text fragments to generate a concise and user - interpretable shopping guide explanation (i.e., characteristic information), which can include information such as the main features of the SPU, applicable scenarios, and user reviews.

[0233] A search device provided by an exemplary embodiment of the present application. The device includes: an acquisition model and a display module. Among them, the acquisition module is used to respond to a search event, acquire the search results related to the keyword, and acquire the reference information for assisting the user output by the preset model based on the search results. The display module is used to display the search results and the reference information on the interface.

[0234] In an implementable example, the reference information is navigation information, and the navigation information includes: multiple classification titles obtained by the preset model analyzing and classifying multiple result items in the search results. Correspondingly, when the display module displays the navigation information, it is specifically configured to: display first operable controls respectively corresponding to the multiple classification titles.

[0235] Furthermore, the navigation information further includes: at least one result item in the classification group corresponding to the classification title. Correspondingly, the display module is further configured to, in response to a trigger of the first operable control corresponding to a target classification title among the multiple classification titles, display page elements of at least some result items in the classification group corresponding to the target classification title; wherein, the page elements of the result item include at least one of a picture of the result item, a video, a difference interval where the same result item provided by different providers has differences summarized by the preset model, and characteristic information of the result item summarized by the preset model.

[0236] Alternatively, the navigation information includes multiple sub-group sets included in the classification group, and each sub-group set includes at least one result item. Correspondingly, the display module is further configured to, in response to a trigger of the first operable control corresponding to a target classification title among the multiple classification titles, display page elements of at least some sub-group sets in the classification group corresponding to the target classification title; wherein, the page elements of the sub-group set include at least one of a picture representing the result items in the sub-group set, a video, a difference interval where the result items included in the sub-group set have differences summarized by the preset model, and characteristic information of the result items included in the sub-group set summarized by the preset model.

[0237] Furthermore, when the display module displays the navigation information, it is configured to display page elements of at least some result items in the classification group corresponding to at least one classification title among the multiple classification titles; wherein, the page elements of the result item include at least one of a picture of the result item, a video, an attribute interval summarized by the preset model, and characteristic information of the result item summarized by the preset model.

[0238] Alternatively, when the display module displays the navigation information, it is configured to display page elements of at least some sub-group sets in the classification group corresponding to at least one classification title among the multiple classification titles; wherein each sub-group set includes at least one result item, and the page elements of the sub-group set include at least one of a picture representing the result items in the sub-group set, a video, a difference interval where the result items included in the sub-group set have differences summarized by the preset model, and characteristic information of the result items included in the sub-group set summarized by the preset model.

[0239] Further, the page elements of the result items are operable. Correspondingly, the display module is further configured to respond to an operation on the page elements of a result item or a subset group set on the interface, and display a drop-down menu, where the drop-down menu displays second operable parts of the result items in the result item or subset group set provided by different providers; wherein, operating the second operable part can trigger a page jump to the first page corresponding to the result item provided by the corresponding provider; or respond to an operation on the page elements of a result item or a subset group set on the interface, and jump to a second page that displays the result items in the result item or subset group set provided by different providers.

[0240] In another implementable example, the reference information is recommendation information, and the recommendation information includes at least one recommended result item selected by the preset model from the search results and recommended texts generated for the recommended result items.

[0241] On this basis, when the display module displays the recommendation information, it is specifically configured to: display the page elements of at least some of the recommended result items in the recommendation information; wherein, the page elements of the recommended result items include: the title of the recommended result item, the recommended text, and the picture or video of the recommended result item.

[0242] Further, when the display module displays the search results and the reference information on the interface, it is specifically configured to: display the reference information in the first area of the interface; display the search results in the second area of the interface; wherein, a search box is further displayed on the interface, and the first area is located between the search box and the second area.

[0243] A search device provided by another exemplary embodiment of the present application. The device includes: an acquisition module and a display module. Wherein, the acquisition module is configured to respond to a search event and acquire search results related to a keyword, where the search results are obtained by searching based on a search intent related to the keyword, and the search intent is generated by a first preset model based on the keyword; acquire reference information output by a second preset model based on the search results. The display module is configured to display the search results and the reference information on the interface; wherein, the first preset model and the second preset model are the same model, or two different models; the reference information is navigation information for assisting the user to understand the search results, or recommendation information; the recommendation information includes at least one recommended result item selected by the preset model from the search results and recommended texts generated for the recommended result items.

[0244] A search device provided by another exemplary embodiment of the present application. The device includes: an acquisition module and a display module. Among them, the acquisition module is used to respond to a search event, acquire a first search result related to a keyword searched by a search engine; acquire a second search result related to the keyword, where the second search result is obtained by the search engine based on a search intent related to the keyword, and the search intent is generated by a first preset model based on the keyword; acquire first reference information output by a second preset model based on the second search result. The display module is used to display the first search result and the first reference information on an interface. Among them, the first preset model and the second preset model are the same model or two different models; the first reference information is first navigation information or first recommendation information for assisting the user to understand the second search result; the first recommendation information includes at least one recommended result item selected by the preset model from the second search result and a recommended text generated for the recommended result item.

[0245] Further, the acquisition module is further used to: acquire second reference information output by the second preset model based on the first search result and the second search result; where the second reference information is second navigation information for assisting the user to understand the first search result and the second search result, or second recommendation information; the second recommendation information includes at least one recommended result item selected by the preset model from the first search result and the second search result and a recommended text generated for the recommended result item. Correspondingly, the display module is further used to: display the first search result, the second search result, and the second reference information on the interface.

[0246] A search device provided by another exemplary embodiment of the present application. The device includes: an acquisition module, a search module, a processing module, and a sending module. Among them, the acquisition module is used to acquire a keyword. The search module is used to search for a search result related to the keyword. The processing module is used to use a preset model to process the search result to obtain reference information for assisting the user. The sending module is used to send the search result and the reference information to a client for display on a client interface.

[0247] Further, the device further includes a generation module, which is used to use the preset model to analyze the keyword to determine the search intent; generate at least one search term based on the search intent. Correspondingly, when the search module searches for a search result related to the keyword, it is specifically used to: use a search engine to search for a first result set related to the keyword; and / or use a search engine to search for a second result set related to the at least one search term; where the search result includes the first result set and / or the second result set.

[0248] Alternatively, when the search module obtains search results related to the keyword, it specifically performs the following operations: using a search engine to search for a first result set related to the keyword; using a preset model to determine the search intent based on the first result set; generating at least one search term based on the search intent; using the search engine to search for a second result set related to the at least one search term; and determining the search results based on the first result set and the second result set.

[0249] In a specific example, the reference information is navigation information. Correspondingly, when the processing module uses the preset model to process the search results to obtain navigation information, it specifically performs the following operations: the preset model analyzes the search results to determine a classification method; and the preset model classifies the search results according to the classification method to obtain multiple classification groups and corresponding classification titles.

[0250] Further, when the processing module uses the preset model to process the search results to obtain navigation information, it is further configured to: the preset model analyzes the result items included in the classification group to summarize the characteristic information of the result items and / or the difference intervals where the same result items provided by different providers differ.

[0251] When the result items in the search results are products, the classification method includes at least one of the following: product brand, product category, and the standardized product unit (SPU) of the product.

[0252] Further, there are multiple alternative classification methods preset. Correspondingly, when the processing module uses the preset model to analyze the search results to determine the classification method, it specifically performs the following operations:

[0253] The preset model selects one from the multiple alternative classification methods by analyzing the search results and based on the analysis results; or

[0254] The preset model selects one from the multiple alternative classification methods according to the search intent and the search results, where the search intent is determined by the preset model based on the keyword.

[0255] Further, the device provided in this embodiment may further include an evaluation module. The evaluation module is used to perform a quality evaluation on multiple classification groups obtained according to the selected classification method. The processing module is used to trigger the preset model to analyze the result items included in the classification group and summarize the characteristic information of the result items and / or the difference interval where the same result items provided by different providers differ when the quality evaluation passes; or group the result items included in the classification group to obtain a subset of sub-classification groups within the classification group, sort the multiple subsets of sub-classification groups within the group, and summarize the characteristic information of the result items within the subset of sub-classification groups and / or the difference interval where the result items within the subset of sub-classification groups differ; when the quality evaluation fails, the preset model selects another one from the multiple alternative classification methods.

[0256] When the multiple classification groups obtained by classifying the search results according to the re-selected classification method still fail the quality evaluation, the processing module in the device provided in this embodiment is further used to: use the search term as the classification method and aggregate the search results corresponding to different search terms into one classification group; where the multiple search terms are determined by the preset model through intention recognition of the keywords.

[0257] In another specific example, the reference information is recommendation information. Correspondingly, when the processing module uses the preset model to process the search results to obtain the recommendation information, it is specifically used to:

[0258] The preset model analyzes the search results, selects at least one recommended result item from the search results, and generates a recommended text for the recommended result item.

[0259] A search device provided by another exemplary embodiment of the present application. The device includes: an acquisition module and a display module. The acquisition module is used to respond to a commodity search event, obtain a commodity search result, and obtain reference information for assisting users in selecting products output by the preset model based on the commodity search result; the display module is used to display the search result and the reference information on the interface.

[0260] It should be noted here that: the devices provided in the above embodiments can implement the technical solutions described in the corresponding method embodiments in the above text. The specific implementation principles and corresponding beneficial effects of the above modules or units can be seen in the corresponding content of the above method embodiments, and will not be repeated here.

[0261] An embodiment of the present application further provides an electronic device. As Figure 10 shown, the electronic device includes a processor 42 and a memory 41. The memory 41 is used to store one or more computer programs; the processor 42 is coupled to the memory 41 and is used for the at least one or more computer programs to implement the steps in the methods provided in the embodiments of the present application.

[0262] The above-mentioned memory 41 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0263] Furthermore, the electronic device further includes other components such as a communication component 43, a display 44, a power supply component 45, and an audio component 46. Only some components are schematically shown here, which does not mean that the electronic device only has these components.

[0264] Correspondingly, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a computer, it can implement the method steps or functions provided in the above embodiments of the present application.

[0265] The method in the present application can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. Thus, the present application also provides a computer program product. The computer program product includes computer programs / instructions. When the computer programs / instructions are executed by an electronic component such as a processor, they can execute all or part of the steps or functions in the methods provided in the embodiments of the present application. The computer can be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user device, a core network device, an OAM, or other programmable devices.

[0266] The above-mentioned memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0267] The above-mentioned display includes a screen, which may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen can be implemented as a touch screen to receive input signals from users. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors can not only sense the boundaries of touch or swipe actions, but also detect the duration and pressure associated with touch or swipe operations.

[0268] The above-mentioned power supply component provides power for various components of the device where the power supply component is located. The power supply component may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device where the power supply component is located.

[0269] The above-mentioned audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC). When the device where the audio component is located is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode, the microphone is configured to receive external audio signals. The received audio signals can be further stored in the memory or sent via the communication component. In some embodiments, the audio component further includes a speaker for outputting audio signals.

[0270] Accordingly, an embodiment of the present application further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it causes the processor to be able to implement the steps in the above method embodiments. Among them, the computer-readable storage medium can be implemented by volatile or non-volatile or a combination thereof, and can be removable or non-removable. Examples of computer-readable storage media include, but are not limited to, Phase-change Random Access Memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), flash memory or other memory technologies, Compact Disc Read Only Memory (CD-ROM), Digital Video Disc (DVD) or other optical storage, magnetic cassette tapes, magnetic disk storage or other magnetic storage devices or any other non-transmission medium

[0271] Accordingly, an embodiment of the present application further provides a computer program product, which includes a computer program or instructions. When the computer program or instructions are executed by a processor, the processor is enabled to implement each step in the above method embodiment. It should be understood that each process or a combination of multiple processes in the above method flow can be implemented by the computer program or instructions. In addition, these computer programs or instructions can be applied to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices, so that the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices can be used as devices to implement the corresponding functions in the above method embodiment.

[0272] It should also be noted that the term "including", "comprising", or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, commodity, or device including a series of elements not only includes those elements but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, commodity, or device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, commodity, or device including the element.

[0273] The above are only embodiments of the present application and are not used to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A search method, characterized in that: include: Respond to search events and obtain search results related to keywords; Obtaining reference information output by a preset model based on the search results to assist the user in understanding the search results; The search results and the reference information are displayed on the interface.

2. The method according to claim 1, characterized in that The reference information is navigation information, and the navigation information includes: a plurality of classification titles obtained by analyzing and classifying a plurality of result items in the search results by the preset model; and Displaying the navigation information includes: Display the first operable controls corresponding to the multiple category titles respectively.

3. The method according to claim 2, characterized in that The navigation information includes: at least one result item in the classification group corresponding to the classification title, and the method further includes: In response to triggering of a first operable control corresponding to a target classification title among the multiple classification titles, display page elements of at least some result items in the classification group corresponding to the target classification title; wherein the page elements of the result items include at least one of pictures and videos of the result items, difference intervals of the same result item provided by different providers summarized by the preset model, and characteristic information of the result item summarized by the preset model; or The navigation information includes multiple sub-grouping sets contained in the classification group, and the sub-grouping set includes at least one result item, and the method also includes: in response to the triggering of a first operable control corresponding to a target classification title among the multiple classification titles, displaying page elements of at least part of the sub-grouping sets in the classification group corresponding to the target classification title; wherein the page elements of the sub-grouping set include pictures and videos representing the result items in the sub-grouping set, difference intervals where the result items contained in the sub-grouping set summarized by the preset model are different, and at least one of the feature information of the result items contained in the sub-grouping set summarized by the preset model.

4. The method according to claim 2, characterized in that: Displaying the navigation information further includes: Displaying page elements of at least some result items in the classification group corresponding to at least one of the multiple classification titles; wherein the page elements of the result items include at least one of the pictures and videos of the result items, the difference intervals of the same result item provided by different providers summarized by the preset model, and the characteristic information of the result item summarized by the preset model; or Display page elements of at least part of the sub-grouping set in the classification group corresponding to at least one classification title among the multiple classification titles; wherein the sub-grouping set includes at least one result item, and the page elements of the sub-grouping set include pictures and videos representing the result items in the sub-grouping set, difference intervals where differences exist in the result items included in the sub-grouping set summarized by the preset model, and at least one item of feature information of the result items included in the sub-grouping set summarized by the preset model.

5. The method according to claim 3 or 4, characterized in that: The page elements of the result item are operable; as well as The method further comprises: In response to an operation on a page element of a result item or a subgrouping set on the interface, a drop-down menu is displayed, wherein the drop-down menu displays a second operable element of the result item or the subgrouping set provided by a different provider; wherein operating the second operable element can trigger a page jump to a first page corresponding to the result item; or In response to an operation on a page element of a result item or a sub-grouping set on the interface, jump to a second page displaying the result item or the result item in the sub-grouping set provided by a different provider.

6. The method according to claim 1, characterized in that The reference information is recommendation information, and the recommendation information includes at least one recommendation result item selected by the preset model from the search results and a recommendation text generated for the recommendation result item.

7. The method according to claim 6, characterized in that Displaying the recommended information includes: A page element showing at least some of the recommended result items in the recommendation information; The page elements of the recommended result item include: the title of the recommended result item, the recommended text, and the picture or video of the recommended result item.

8. The method according to claim 1, characterized in that The search results and reference information are displayed on the interface, including: The reference information is displayed in the first area of ​​the interface; The search results are displayed in a second area of ​​the interface; A search box is also displayed on the interface, and the first area is located between the search box and the second area.

9. A search method, characterized in that: include: In response to a search event, obtaining search results related to the keyword, wherein the search results are obtained based on a search intent related to the keyword, and the search intent is generated by a first preset model based on the keyword; Acquire reference information output by a second preset model based on the search results; On the interface, display the search results and the reference information; Among them, the first preset model and the second preset model are the same model, or two different models; the reference information is navigation information used to assist users in understanding search results or is recommendation information; the recommendation information includes at least one recommended result item selected by the second preset model in the search results and a recommendation text generated for the recommended result item.

10. A search method, characterized in that: include: In response to the search event, obtaining the first search result related to the keyword searched by the search engine; Obtaining a second search result related to the keyword, where the second search result is obtained by searching the search engine based on a search intent related to the keyword, where the search intent is generated by the first preset model based on the keyword; Acquire first reference information output by a second preset model based on the second search result; On the interface, displaying the first search result and the first reference information; The first preset model and the second preset model are the same model or two different models; the first reference information is first navigation information or first recommendation information for assisting the user to understand the second search result; The first recommendation information includes at least one recommendation result item selected by the second preset model from the second search results and a recommendation text generated for the recommendation result item.

11. The method according to claim 10, characterized in that Acquiring first reference information output by a second preset model based on the second search result includes: Obtaining second reference information output by the second preset model based on the first search result and the second search result; wherein the second reference information is second navigation information for assisting the user to understand the first search result and the second search result, or is second recommendation information; the second recommendation information includes at least one recommended result item selected by the preset model from the first search result and the second search result and a recommendation text generated for the recommended result item; and Displaying the first search result and the first reference information on an interface includes: displaying the first search result, the second search result and the second reference information on the interface.

12. A search method, characterized in that: include: Get keywords; Search to obtain search results related to the keyword; Using a preset model, the search results are processed to obtain reference information that assists the user in understanding the search results; The search results and the reference information are sent to the client to be displayed on the client interface.

13. The method according to claim 12, characterized in that Also includes: Analyze the keywords using the preset model to determine the search intent; generating at least one search term based on the search intent; And the searching to obtain search results related to the keyword includes: using a search engine to search for a first result set related to the keyword; and / or using a search engine to search for a second result set related to the at least one search term; The search results include the first result set and / or the second result set.

14. The method according to claim 12, characterized in that Search results related to the keyword include: Using a search engine, searching for a first result set related to the keyword; Determine the search intent based on the first result set using a preset model; generating at least one search term based on the search intent; Using a search engine, searching for a second result set related to the at least one search term; The search result is determined according to the first result set and the second result set.

15. The method according to any one of claims 12 to 14, characterized in that The reference information is navigation information; The search results are processed using a preset model to obtain navigation information, including: The preset model analyzes the search results to determine a classification method; The preset model classifies the search results according to the classification method to obtain multiple classification groups and classification titles corresponding to the classification groups.

16. The method according to claim 15, characterized in that Using a preset model, processing the search results to obtain navigation information also includes: The preset model analyzes the result items included in the classification group, summarizes the characteristic information of the result items and / or the difference intervals of the same result item provided by different providers.

17. The method according to claim 15, characterized in that When the result item in the search result is a commodity, the classification method includes at least one of the following: commodity brand, commodity category and standardized product unit SPU of the commodity.

18. The method according to claim 16, characterized in that There are multiple preset classification options, and The preset model analyzes the search results to determine a classification method, including: The preset model analyzes the search results and selects one of the multiple candidate classification methods based on the analysis results; or The preset model selects one from the multiple alternative classification methods according to the search intent and the search results; wherein the search intent is determined by the preset model based on the keywords.

19. The method according to claim 18, characterized in that Also includes: performing a quality assessment on the plurality of classification groups; When the quality assessment is passed, the preset model is triggered to analyze the result items included in the classification group, summarize the characteristic information of the result items and / or the difference intervals of the same result items provided by different providers; or the result items included in the classification group are grouped to obtain sub-grouping sets within the classification group, multiple sub-grouping sets within the group are sorted, and the characteristic information of the result items in the sub-grouping sets and / or the difference intervals of the result items in the sub-grouping sets are summarized; When the quality assessment fails, another one of the multiple candidate classification methods is selected.

20. The method according to claim 19, characterized in that When the multiple classification groups obtained by classifying the search results according to the classification method selected again still fail to pass the quality assessment, the method further includes: Search terms are used as a classification method to aggregate search results corresponding to different search terms into a classification group; wherein, multiple search terms are determined by a preset model through intent recognition of keywords, or are determined by a preset model based on a first result set; the first result set is searched by the search engine based on the keywords.

21. The method according to any one of claims 12 to 14, characterized in that The reference information is recommended information; The search results are processed using a preset model to obtain recommendation information, including: The preset model analyzes the search results, selects at least one recommended result item from the search results, and generates a recommended text for the recommended result item.

22. A commodity search method, characterized in that: include: Respond to product search events and obtain product search results; Obtaining a preset model to output reference information to assist users in selecting products based on the product search results; The search results and the reference information are displayed on the interface.

23. A search system, characterized in that: include: A client, used to execute the steps of the search method described in any one of claims 1 to 11; A server, used to execute the steps in the search method described in any one of claims 12 to 21.

24. An electronic device, characterized in that: Memory and processor; wherein, The memory is used to store executable instructions; The processor implements the steps in the search method according to any one of claims 1 to 21 or the steps in the commodity search method according to claim 22 by running the executable instructions.

25. A computer-readable storage medium, characterized in that: The storage medium stores computer instructions, which, when executed by a processor, can implement the steps in the search method described in any one of claims 1 to 21, or the steps in the commodity search method described in claim 22.

26. A computer program product, characterized in that The computer program product includes a computer program or instructions, which, when executed by a processor, enables the processor to execute the steps in the search method described in any one of claims 1 to 21, or the steps in the commodity search method described in claim 22.

Images