Data processing method and device, equipment and medium

Abstract

The embodiment of the invention provides a data processing method and device, equipment and a medium, and the method comprises the steps: responding to a search operation for query information, and displaying a first result page corresponding to the query information; in the first result page, displaying a search result corresponding to the query information and a search extension floating layer associated with the query information; the search extension floating layer is independently displayed on a first result page, and a tag page column of the first result page comprises a search extension tag; and in response to a detail viewing operation for the search extension floating layer, displaying a concerned topic directory associated with the query information in a channel page corresponding to the search extension tag. By implementing the embodiment of the invention, the search traffic can be improved.

Description

Technical Field

[0001] The present application relates to the field of computer network technology, and in particular to a data processing method, device, equipment and medium. Background Art

[0002] In the current search scenario, when an object (for example, object A) enters a query term in a search engine and performs a search operation on the query term, the search engine can obtain the query term entered by object A; perform information retrieval based on the query term, obtain the search results corresponding to the query term, and present the search results to object A in a certain order and format.

[0003] At the same time, in addition to presenting the search results corresponding to the query term to object A, the search needs of object A can also be analyzed through the query term, and recommended terms associated with the query term can be presented to object A. Object A can click on any recommended term that he is interested in to view it. However, for all the recommended terms associated with the query term, object A usually only selects one recommended term to view, resulting in a too low click-through rate (CTR, which can be understood here as the ratio of the number of times a recommended term is clicked to the number of times it is displayed) of these recommended terms, reducing search traffic. Summary of the invention

[0004] The embodiments of the present application provide a data processing method, apparatus, device and medium, which can improve search traffic.

[0005] On the one hand, an embodiment of the present application provides a data processing method, including:

[0006] In response to a search operation for the query information, display a first result page corresponding to the query information;

[0007] On the first result page, the search results corresponding to the query information and the search expansion floating layer associated with the query information are displayed; the search expansion floating layer is displayed independently on the first result page, and the tab bar of the first result page includes a search expansion tag;

[0008] In response to a detail viewing operation on the search extension floating layer, a directory of topics of interest associated with the query information is displayed in a channel page corresponding to the search extension tag.

[0009] An embodiment of the present application provides a data processing device, including:

[0010] A page display module, used to respond to a search operation for query information and display a first result page corresponding to the query information;

[0011] The page display module is further used to display the search results corresponding to the query information and the search expansion floating layer associated with the query information in the first result page; the search expansion floating layer is independently displayed on the first result page, and the tab bar of the first result page includes the search expansion tag;

[0012] The directory display module is used to respond to the detail viewing operation on the search extension floating layer, and display the directory of concerned topics associated with the query information in the channel page corresponding to the search extension tag.

[0013] The page display module is also used to perform the following steps:

[0014] In the entry area of ​​the search extension floating layer, the recommended entries of the topic of interest associated with the query information are displayed in a carousel;

[0015] In response to a trigger operation on the entry area, determining the recommended entry displayed in the entry area as a target entry, and displaying a second result page corresponding to the target entry;

[0016] In the second result page, the search results and directory viewing controls corresponding to the target term are displayed; the search extension tab in the tab bar of the second result page is in a hidden state.

[0017] The page display module is also used to perform the following steps:

[0018] In response to a trigger operation on a directory viewing control in the second result page, a directory of topics of interest associated with the query information is displayed in a channel page corresponding to the search extension tag;

[0019] The target entry in the focused topic catalog is highlighted, and in the focused topic catalog, an extended entry is added to the focused topic to which the target entry belongs.

[0020] The page display module adds an extended entry to the concerned topic to which the target entry belongs in the concerned topic directory, so as to perform the following steps:

[0021] According to the display position of the target term in the channel page corresponding to the search extension tag, the term connection animation is displayed in the focus topic directory; the term connection animation is used to rotate the extended terms of the focus topic to which the target term belongs;

[0022] When the entry connection animation is finished, the extended entries included in the entry connection animation are expanded and displayed in the focus topic directory, and the display positions of multiple focus topics in the focus topic directory and the recommended entries of the multiple focus topics are adjusted.

[0023] The directory display module responds to the detail viewing operation on the search extension floating layer, and displays the directory of concerned topics associated with the query information in the channel page corresponding to the search extension tag, and is used to perform the following steps:

[0024] In response to a detail viewing operation on the search expansion floating layer, multiple topics of interest associated with the query information and recommended terms corresponding to each topic of interest are obtained;

[0025] Determine multiple topics of interest associated with the query information as first-level titles, and determine recommended terms corresponding to each topic of interest as second-level titles;

[0026] According to the primary title and the secondary title, a directory of topics of interest associated with the query information is constructed, and the directory of topics of interest is displayed in a channel page corresponding to the search extension tag.

[0027] The page display module is also used to perform the following steps:

[0028] In response to a triggering operation on a search extension tag in the first result page, displaying a channel page corresponding to the search extension tag;

[0029] In the channel page corresponding to the search extension tag, a directory of topics of interest and a result return control associated with the query information are displayed;

[0030] The focus topic directory includes multiple focus topics associated with the query information and recommended terms corresponding to the multiple focus topics; the result return control is used to return from the channel page corresponding to the search extension page to the first result page.

[0031] The page display module is also used to perform the following steps:

[0032] In response to a sliding operation on the first result page, updating displayed content on the first result page according to a sliding speed indicated by the sliding operation;

[0033] When the displayed content in the first result page includes the focus display area, hiding the displayed search expansion floating layer in the first result page;

[0034] The focus display area is used to display multiple focus topics associated with the query information and recommended terms corresponding to the multiple focus topics.

[0035] The focus display area also includes a detail viewing control;

[0036] The page display module is also used to perform the following steps:

[0037] In response to a trigger operation on a detail viewing control in the focus display area, a channel page corresponding to the search extension tag in the first result page is displayed, and a directory of focus topics associated with the query information is displayed in the channel page corresponding to the search extension tag;

[0038] The focus topic catalog includes multiple focus topics contained in the focus point display area and recommended entries corresponding to the multiple focus topics.

[0039] The page display module is also used to perform the following steps:

[0040] If it is detected that among the recommended terms of the plurality of focus topics associated with the query information, there are recommended terms that have been viewed, then in the focus point display area, the recommended terms that have been viewed and the recommended terms that have not been viewed are displayed separately.

[0041] The data processing device further comprises:

[0042] A retrieval module is used to obtain query information and search the query information in an index database; the index database includes multiple index information, one index information corresponds to multiple focus topics, and one focus topic includes multiple recommended candidate entries;

[0043] The retrieval module is further configured to obtain, in the index database, a plurality of topics of interest corresponding to the target index information and a plurality of recommended candidate entries for each topic of interest corresponding to the target index information if target index information matching the query information is retrieved in the index database;

[0044] A term screening module, used to screen multiple recommended candidate terms for each topic of interest corresponding to the target index information, and obtain a recommended term for each topic of interest corresponding to the target search information;

[0045] The directory generation module is used to generate a directory of topics of interest associated with the query information according to multiple topics of interest corresponding to the target index information and the recommended terms of each topic of interest corresponding to the target search information.

[0046] Among them, the index database includes an inverted index database and a semantic index database;

[0047] The retrieval module obtains the query information and searches the query information in the index database to perform the following steps:

[0048] Obtain query information, perform semantic analysis on the query information, and obtain a text semantic vector corresponding to the query information;

[0049] If a query statement matching the query information is retrieved in the inverted index library, the query statement matching the query information in the inverted index library is determined as target index information of the query information in the inverted index library;

[0050] If a semantic index matching the text semantic vector is retrieved in the semantic index library, the semantic index matching the text semantic vector in the semantic index library is determined as the target index information of the query information in the semantic index library.

[0051] The entry screening module is specifically used to perform the following steps:

[0052] Obtaining a text semantic vector corresponding to the query information, and adding a plurality of recommended candidate words for each topic of interest corresponding to the target index information to the first candidate word set;

[0053] Obtaining the term semantic vector corresponding to each recommended candidate term in the first candidate term set, and obtaining the first similarity between the text semantic vector and the term semantic vector;

[0054] In the first candidate term set, the recommended candidate terms whose first similarity is less than the first similarity threshold are deleted to obtain a second candidate term set;

[0055] Obtaining a feedback index estimate value corresponding to each recommended candidate term in the second candidate term set, and obtaining a second similarity between each recommended candidate term in the second candidate term set according to the term semantic vector;

[0056] If in the second candidate term set, there exists a second similarity between the recommended candidate term a and the recommended candidate term b that is greater than the second similarity threshold, and the feedback index estimated value of the recommended candidate term a is greater than the feedback index estimated value of the recommended candidate term b, then the recommended candidate term b is deleted from the second candidate term set to obtain a third candidate term set;

[0057] Among the recommended candidate terms included in the third candidate term set, a recommended term for each topic of interest corresponding to the target search information is determined.

[0058] The directory generation module is specifically used to perform the following steps:

[0059] Determine the estimated feedback index value of each topic of interest corresponding to the target index information according to the estimated feedback index value corresponding to the recommended term of each topic of interest;

[0060] According to the estimated value of the feedback indicator of each topic of interest, multiple topics of interest corresponding to the target index information are sorted to obtain multiple sorted topics of interest;

[0061] According to the estimated feedback indicator value corresponding to each recommended term of the concerned topic, the recommended terms of each concerned topic are sorted to obtain the sorted recommended terms corresponding to each concerned topic;

[0062] A directory of topics of interest associated with the query information is generated according to the sorted multiple topics of interest and the sorted recommended terms corresponding to each topic of interest.

[0063] The data processing device further comprises:

[0064] A query statement collection module is used to obtain historical query statements and M historical candidate entries corresponding to the historical query statements; M is an integer greater than 1;

[0065] A focus generation module, used to input the historical query statement into the language model, generate a first branch topic corresponding to the historical query statement through the language model, and generate a first branch term corresponding to the first branch topic through the language model;

[0066] The term clustering module is used to perform clustering processing on the M historical candidate terms corresponding to the historical query sentence to obtain N candidate term groups, and generate a second branch topic for each candidate term group; N is a positive integer;

[0067] A topic deduplication module is used to perform deduplication processing on the first branch terms and the second branch topics to obtain the concerned topics corresponding to the historical query statements;

[0068] A term deduplication module is used to perform deduplication processing on the M historical candidate terms and the first branch terms corresponding to the first branch topic, so as to obtain the recommended candidate terms of the concerned topic corresponding to the historical query statement;

[0069] The adding module is used to add the historical query statements, the concerned topics corresponding to the historical query statements, and the recommended candidate entries of the concerned topics corresponding to the historical query statements to the index database.

[0070] The focus generation module inputs the historical query statement into the language model, and generates a first branch topic corresponding to the historical query statement through the language model, which is used to perform the following steps:

[0071] Input the historical query sentences into the language model, and convert the historical query sentences into historical embedding representations through the embedding layer in the language model;

[0072] Through the encoding component in the language model, the historical embedding representation is encoded to obtain the historical encoding features corresponding to the historical query sentences;

[0073] Through the fully connected layer in the language model, the historical encoding features are transformed to generate the first branch topic corresponding to the historical query statement.

[0074] The term clustering module performs clustering processing on the M historical candidate terms corresponding to the historical query statements to obtain N candidate term groups, and generates a second branch topic for each candidate term group to perform the following steps:

[0075] Convert the M historical candidate terms corresponding to the historical query statements into historical text vectors, add the historical text vectors corresponding to the M historical candidate terms to the text vector set, and select the center vector t in the text vector set. i ; i is a positive integer less than or equal to N;

[0076] Get each historical text vector and the center vector t in the text vector set i The third similarity between them, if the historical text vector G in the text vector set j With the center vector t i The third similarity between them is the largest, so the historical text vector G j Add to the center vector t i The pending cluster C i ; j is a positive integer less than or equal to M;

[0077] According to the cluster C to be processed i The historical text vector contained in the update center vector t i , until the cluster to be processed C i The updated center vector t i and the center vector t before update i If the same, according to the cluster C to be processed i Obtain N candidate entry groups;

[0078] The historical candidate terms in each candidate term group are input into the language model, and the second branch topic corresponding to each candidate term group is generated through the language model.

[0079] On one hand, an embodiment of the present application provides a computer device, including a memory and a processor, wherein the memory is connected to the processor, the memory is used to store a computer program, and the processor is used to call the computer program so that the computer device executes the method provided in the above aspect of the embodiment of the present application.

[0080] On one hand, an embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored. The computer program is suitable for being loaded and executed by a processor, so that a computer device with a processor executes the method provided in the above aspect of the embodiment of the present application.

[0081] According to one aspect of the present application, a computer program product is provided, which may include a computer program stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the method provided in the above aspect.

[0082] In an embodiment of the present application, when an object enters query information in a search box, a search operation for the query information can be responded to, and a first result page corresponding to the query information can be displayed, and the search results and a search expansion floating layer corresponding to the query information can be displayed in the first result page. The search expansion floating layer is displayed independently on the first result page, and the tab bar of the first result page can include a search expansion tag; both the search expansion floating layer and the search expansion tag can be used to jump to the channel page corresponding to the search expansion tag, and the directory of topics of interest associated with the query information can be displayed in the channel page corresponding to the search expansion tag. Since the directory of topics of interest contains multiple topics of interest associated with the query information, as well as recommended terms under each topic of interest, and the directory of topics of interest is displayed in the channel page corresponding to the search expansion tag, the object can click on the recommended terms in the topic directory to view them in turn, which can improve the click-through rate of the directory of topics of interest, and thus increase search traffic. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0084] Figure 1 It is a structural diagram of a network architecture provided by an embodiment of the present application;

[0085] Figure 2 This is a flow diagram of a data processing method provided in an embodiment of the present application. Figure 1 ;

[0086] Figure 3 This is a schematic diagram of an interface of a floating layer entrance in a result page provided in an embodiment of the present application. Figure 1 ;

[0087] Figure 4 This is a schematic diagram of an interface of a floating layer entrance in a result page provided in an embodiment of the present application. Figure 2 ;

[0088] Figure 5 This is a schematic diagram of an interface of a concerned topic directory provided in an embodiment of the present application;

[0089] Figure 6 This is a schematic diagram of an interface of a tab entry in a tab bar of a result page provided in an embodiment of the present application;

[0090] Figure 7 This is a schematic diagram of an interface in which a directory of topics of interest fails to load, provided in an embodiment of the present application;

[0091] Figure 8 is a schematic diagram of an interface of a search process provided in an embodiment of the present application;

[0092] Figure 9a This is a schematic diagram of an interface of an entrance to a focus display area in a result page provided in an embodiment of the present application. Figure 1 ;

[0093] Figure 9b This is a schematic diagram of an interface of an entrance to a focus display area in a result page provided in an embodiment of the present application. Figure 2 ;

[0094] Fig.10 This is a flow diagram of a data processing method provided in an embodiment of the present application. Figure 2 ;

[0095] Fig.11 It is a schematic diagram of an architecture strategy for a related search scenario provided in an embodiment of the present application;

[0096] Fig.12 is a structural schematic diagram of a data processing device provided in an embodiment of the present application;

[0097] Fig.13 It is a structural diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0098] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0099] See also Figure 1 , Figure 1 1 is a schematic diagram of a network architecture provided by an embodiment of the present application; the network architecture may include a server 10d and a terminal cluster, the terminal cluster may include one or more terminal devices, and the number of terminal devices included in the terminal cluster is not limited. Figure 1As shown, the terminal cluster may specifically include terminal device 10a, terminal device 10b, and terminal device 10c, etc.; all terminal devices in the terminal cluster (for example, including terminal device 10a, terminal device 10b, and terminal device 10c, etc.) may be connected to the server 10d through a network connection, so that each terminal device may exchange data with the server 10d through the network connection.

[0100] in, Figure 1 The terminal devices in the terminal cluster shown may include but are not limited to: smart phones, tablet computers, laptops, PDAs, desktop computers, wearable devices (such as smart watches, smart bracelets, etc.), smart voice interaction devices, smart home appliances (such as smart TVs, etc.), vehicle-mounted devices, aircraft and other electronic devices. This application does not limit the type of terminal devices.

[0101] Figure 1 The server 10d shown can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms. This application does not limit the type of server.

[0102] In the embodiment of the present application, Figure 1 Each terminal device in the terminal cluster shown can run a browser or client. Among them, the browser can be used to access web pages, interact with the server 10d through protocols such as HTTP (HyperText Transfer Protocol) to obtain web page content for presentation to users; the browser usually does not need to be installed or configured for specific services or applications. The client is usually developed for a specific application or service, and usually needs to be downloaded, installed and updated to adapt to the needs and changes of the specific service; the client involved in this application can be any client with a search function, and this application does not limit the type of client. When the client runs in each terminal device, it can communicate with Figure 1 The server 10d shown in the figure performs data interaction, and the server 10d at this time may be a background server corresponding to the client. Among them, the client running in each terminal device may be an independent client or an embedded sub-client integrated in an independent client, and this application does not limit this.

[0103] In the Related Searches (RS) scenario, objects can be found by Figure 1 The search function of the client or browser running in any terminal device (for example, terminal device 10a) shown in the figure queries information that meets its own needs; that is, the essence of search can be considered as an efficient process of obtaining satisfactory answers. Generally, an object can use the search function of the client or browser to perform a "multi-focus direction change word" search operation around a certain search target to obtain relevant information that meets its own needs; wherein the search target may refer to the information that the object actually needs to query; "multi-focus direction change word" may refer to the operation that the object uses different query words to perform multiple searches around a certain search target. Related search may refer to that in a search engine, when an object enters a query word, the search engine displays other query words or short words that are associated or similar to the query word in the result page corresponding to the query word; these other query words or phrases that are associated or similar to the current query word can be called recommended terms. A search engine is a tool or system for finding information on the Internet, and the search function of a browser or client is implemented by integrating a search engine.

[0104] For example, around a certain search target, the object can use related query words to perform scattered back and forth searches, such as the object's search path can be "query word A-query word B-query word C-query word D-query word C-query word A-query word C-query word D", where query word A, query word B, query word C, and query word D are all related terms, such as different descriptive words with the same meaning, different symptoms of a certain disease, etc. Optionally, around a certain search target, the object can also search by repeatedly enlarging / reducing the scope of the query word, such as the object's search path can be "buying a house in xx city-demobilization resettlement policy-demobilization resettlement policy in xx city". Optionally, the object can also search by searching for the preconditions of its own needs in steps, or by continuous search of the same type of objects (people or things).

[0105] In short, when the object uses the search function of the client or browser to query information, it may not be possible to obtain all the information that meets its needs through one query word, and it is necessary to constantly change the query word for multiple searches. These changed query words can be implemented in the form of recommended terms; for example, when the object enters query word A in the client or browser running on the terminal device 10a for search, the search results of the query word A and some recommended terms associated with the query word A can be displayed in the result page of the query word A. The object can select the recommended terms of interest from these recommended terms and click to view them. At this time, the recommended terms selected by the object can be called query word B. In the result page of query word B, the search results of the query word B and the recommended terms associated with the query word B can also be displayed. Optionally, if the object is not interested in some of the recommended terms associated with the query word A, the query word B can be re-entered in the browser or client for search. It is understandable that the object can continue to select recommended terms associated with query term B for search, or can choose to stop searching; when the object stops searching, the series of search operations of the object in the browser or client can be called session (session control); session can be used to record the state of the object, and session refers to a series of interactions between a single client and a server over a period of time.

[0106] In the embodiment of the present application, the recommendation module in the related search can be organized by constructing a recommendation flow. By establishing a clear "outline directory" perception, it is helpful to help the object view the recommended terms of interest, and the search traffic in the related search can be improved. Among them, the recommendation flow refers to a way of presenting information, according to the object's interest tags, historical search records, etc., the algorithm automatically filters and pushes relevant content to the object. The recommendation module can refer to a part of a system or platform (for example, a browser or client with a search function) used to display personalized recommended content to the object; for example, in a related search scenario, the recommendation module can be used to display the recommended terms associated with the current query term to the object. "Outline directory" can refer to organizing all recommended terms associated with the current query term into a directory of topics of interest, and presenting the directory of topics of interest to the object. The directory of topics of interest can include multiple topics of interest associated with the current query term, and the recommended terms under each topic of interest; topics of interest can refer to topics that are widely concerned, discussed and have a high influence in a specific period or range; recommended terms can be considered as words or phrases recommended for the object; a topic of interest can include one or more recommended terms.

[0107] For the concerned topic directory presented to the object, the object can click in the concerned topic directory to view the recommended terms that he is interested in; for example, after the object clicks on a recommended term in the concerned topic directory, he can view the result of the recommended term, and then return to the concerned topic directory to continue to click to view other recommended terms that he is interested in. The embodiment of the present application adopts the concerned topic directory method, which is conducive to the object repeatedly clicking to view the recommended terms associated with the current query term, thereby improving the click-through rate of these recommended terms and the efficiency of satisfying the needs of the object, thereby improving the search traffic.

[0108] See also Figure 2 , Figure 2 This is a flow diagram of a data processing method provided in an embodiment of the present application. Figure 1 It can be understood that the data processing method can be executed by a computer device, which can be a terminal device, such as Figure 1 Any terminal device in the terminal cluster shown, or a server, such as Figure 1 The server 10d shown in the figure is not limited in this application. Figure 2 As shown, the data processing method may include the following steps S101 to S103:

[0109] Step S101, in response to a search operation for query information, a first result page corresponding to the query information is displayed.

[0110] Step S102, displaying the search results corresponding to the query information and the search expansion floating layer associated with the query information on the first result page; the search expansion floating layer is independently displayed on the first result page, and the tab bar of the first result page includes a search expansion tag.

[0111] The query information may refer to information that an object inputs into a search engine for querying specific information. The query information may be text (words, phrases, sentences, etc.), or may be a picture, or may be audio. This application does not limit the data type of the query information. The first result page may refer to a page for displaying the search results of the query information and related information thereof; the search results may refer to the content related to the query information found by the search engine after analyzing and matching the query information input by the object.

[0112] The search expansion floating layer may refer to a temporary, interactive layer covering the original content in the first result page, and the search expansion floating layer may be displayed independently on the first result page; for example, the search expansion floating layer may be fixedly displayed in the bottom area of ​​the first result page, and the search expansion floating layer at this time may be called the bottom-absorbing floating layer in the first result page; or it may be fixedly displayed at any position in the first result page, and the present application does not limit the display position of the search expansion floating layer. In the search expansion floating layer, multiple recommended terms associated with the query information may be played in turn, and the search expansion floating layer may always be presented in the first result page in the same form; optionally, when the display time of the search expansion floating layer in the first result page is greater than the time threshold (which may be pre-set, such as 5 minutes or any other time, and the present application does not limit this), the search expansion floating layer may be minimized to an icon, and when the object clicks the icon, the original form of the search expansion floating layer may be restored, such as continuing to play the recommended terms associated with the query information in turn, and the present application does not limit the presentation form of the search expansion floating layer in the first result page.

[0113] The search expansion pop-up layer can serve as an entry for the object to view the directory of topics of interest associated with the query information, that is, the object can view the directory of topics of interest associated with the query information through the search expansion pop-up layer. Among them, the directory of topics of interest can be composed of multiple topics of interest associated with the query information, and one or more recommended terms under each topic of interest; all recommended terms contained in the directory of topics of interest are words or phrases displayed to the object based on the query information, and these displayed words or phrases are usually associated with the query information input by the object. The method of determining the topics of interest and recommended terms in the directory of topics of interest will be described in the subsequent steps.

[0114] The tab bar may refer to an area in the first result page for displaying and switching different tabs; the tab bar may include different tabs (or options), and each tab in the tab bar generally represents a different page, document, or functional interface. An object may quickly switch to the page or functional area corresponding to the tab it clicked by clicking on different tabs in the tab bar; for ease of understanding, the embodiment of the present application may refer to the page or functional area corresponding to each tab in the tab bar as a channel page (or tab). The tab bar of the first result page may include a search extension tag, and the channel page corresponding to the search extension tag may be used to display a directory of topics of interest associated with the query information.

[0115] Specifically, when a client or browser with a search function is integrated into a computer device, it means that the client or browser can call a search engine to find content related to the query information input by the object, and present the found content to the object in a certain order and format (which can be pre-set during the development of the result page). For ease of understanding, the embodiment of the present application takes a browser as an example to describe the search process of query information.

[0116] When the object performs a startup operation on the browser in the computer device, the computer device can respond to the startup operation on the browser and display the search page in the browser; the search page can be any interface with a search function, and the search page can include a search area, and a search control can be displayed in the search area. When the object enters information in the search area of ​​the search page, the computer device can respond to the input operation on the search area and obtain the query information entered by the object. After the object completes the input of the query information, it can perform a trigger operation on the search control (the trigger operation here can be understood as a search operation on the query information). At this time, the computer device can respond to the search operation on the query information, call the search engine to analyze and match the query information, and obtain the search results matching the query information, and the search results can be displayed in the first result page corresponding to the query information.

[0117] When the query information input by the object is text, the text input by the object can be displayed in the search area of ​​the search page; for ease of understanding, the text input by the object in the search area can be called the query text. In response to the search operation for the query text, the search engine is called to search for content matching the query text in the pre-established text index; based on the matching degree and algorithm, the text fragments or documents that may meet the object's needs are screened out as search results, and the search results are displayed in the first result page.

[0118] When the query information input by the object is a picture, the object can input the picture to be queried through the picture entrance in the search page (for example, presented in the form of a "camera" icon in the search page); the picture can be a photo taken in real time, or it can be a picture selected from a local gallery, or it can be a picture downloaded from the Internet, etc. This application does not limit the source of the picture; for ease of understanding, the picture input by the object through the picture entrance can be called a query picture. In response to the search operation for the query picture, the content of the query picture can be understood to obtain the translation text corresponding to the query picture; then the search engine can be called to search the translation text, and the search process can refer to the search process of the query text mentioned above, which will not be repeated here.

[0119] When the query information input by the object is audio, the object can input the audio to be queried through the audio entry in the search page (for example, presented in the form of a "microphone" icon in the search page); the audio can be real-time collected speech, or it can be a recording selected from a local database, etc. This application does not limit the source of the audio; for ease of understanding, the audio input by the object through the audio entry can be called query audio. In response to the search operation for the query audio, the query audio can be converted into text data; then the search engine can be called to search the text data, and the search process can refer to the search process of the query text above, which will not be repeated here.

[0120] In an embodiment of the present application, a tab bar may be set in the first result page, and the tab bar may include two or more tabs, each of which may correspond to a channel page; the position of each tab in the tab bar may be pre-set during the development process. Among them, a search extension tab may be set in the tab bar of the first result page, and the position of the search extension tab in the tab bar may be set to the 4th position, or may be set to other positions. The present application does not limit the position of each tab in the tab bar.

[0121] A search expansion floating layer and a focus display area can also be set in the first result page. The search expansion floating layer and the focus display area can be displayed (exposed) at the same time in the first result page, or they can be displayed at different times. This application does not limit this. The focus display area can refer to a card area pre-set in the first result page for displaying focus topics and recommended terms. The card here can refer to an information display unit with a specific layout and presentation method in the first result page. The display position of the focus display area in the first result page can be pre-set according to the sorting strategy of the first result page. For example, the display position of the focus display area can be set to the RS+2 position in the first result page, or it can be set to the RS+1 position in the first result page. This application does not limit this.

[0122] Among them, the sorting strategy of the first result page may refer to a set of rules and methods for determining the order of arrangement of various elements (such as content, information, items, etc.) in the first result page. For example, the elements in the first result page may be sorted according to the relevance of the query information, and the higher the relevance, the higher the display position; or the elements in the first result page may be sorted according to the importance, and the more important the element, the higher the corresponding display position, and so on. The RS position in the first result page may be understood as the exposure position of the first result page in the initial state, and the exposure position here may refer to the area displayed before the interactive operation is performed on the first result page. For example, after the browser enters the first result page, the page area that is first displayed to the object may be called the exposure position of the first result page in the initial state. The RS+1 position and the RS+2 position may be considered as the non-exposed position of the first result page in the initial state, and the non-exposed position here may be understood as the area that needs to be displayed by performing some interactive operations with the first result page (for example, sliding the first result page); the element at the RS+1 position is displayed in the first result page earlier than the element at the RS+2 position. For example, after the browser enters the first result page, the content initially displayed in the first result page does not include the focus display area, and the object needs to slide the first result page to display the focus display area.

[0123] In a feasible implementation, when the search expansion floating layer and the focus display area in the first result page exist at the same time but are not exposed at the same time, the display timing of the search expansion floating layer and the focus display area may include but is not limited to: when the content of the exposure position of the first result page in the initial state is completely exposed, the search expansion floating layer can be continuously displayed in the first result page; when the object slides the first result page until the focus display area is exposed, the search expansion floating layer in the first result page can be hidden and no longer displayed. In other words, before the focus display area is not exposed, the search expansion floating layer can be continuously displayed in the first result page, and after the focus display area is exposed, the search expansion floating layer can be hidden in the first result page.

[0124] If the object continues to slide the first result page, then regardless of whether the current display content of the first result page contains the focus display area, the search expansion floating layer will no longer be displayed. This can ensure the flow of other content in the first result page except the focus data, while also improving the search flow of the focus data. Among them, the focus data involved in the present application may refer to the content used to be displayed in the focus display area, and specifically may include the focus topics retrieved according to the query information, and the recommended terms corresponding to each focus topic. Optionally, after the focus display area is exposed, if the object continues to slide the first result page, then when the focus display area is not included in the current display content of the first result page, the search expansion floating layer can be redisplayed; that is, the focus display area and the search expansion floating layer can be displayed alternately in the first result page.

[0125] In a feasible implementation, when the search expansion floating layer and the focus display area in the first result page exist at the same time and can be exposed at the same time, the search expansion floating layer can be continuously displayed in the first result page; when the object slides the first result page until the focus display area is exposed, the search expansion floating layer in the first result page can be minimized to an icon and displayed in the first result page in the form of an icon. In other words, when the focus display area in the first result page is exposed, that is, the focus display area is displayed in the first result page, the search expansion floating layer can be minimized; when the current display content of the first result display page does not include the focus display area, it can be restored to its original form for display in the first result page. The present application does not limit the display timing of the search expansion floating layer and the focus display area in the first result page. For ease of understanding, the present application is described by taking the example of the search expansion floating layer and the focus display area not being displayed at the same time.

[0126] It is understandable that in the process of searching by inputting query information, retrieval can be performed based on the query information; if the focus topics associated with the query information and the recommended terms corresponding to each focus topic are retrieved, and the number of focus topics and the number of recommended terms both reach a preset first quantity threshold (which can be set according to the actual application scenario, such as 2, which is not limited in this application), it indicates that the current query information can return focus point data. In this case, a search extension tab can be displayed in the tab bar of the first result page, and the search results corresponding to the query result can be displayed in the first result page in a preset specific format, such as displaying the search results corresponding to the query information in the form of a card; at this time, the first result page can also display a search extension floating layer, that is, the search extension tab and the search extension floating layer can be displayed simultaneously in the first result page.

[0127] If no focus topic associated with the query information is retrieved, it means that the current query information cannot return focus data. At this time, the search extension tab and search extension pop-up layer will not be displayed in the tab bar of the first result page, and the focus display area in the first result page will not be displayed; the search results corresponding to the query information can be displayed according to the regular format of the page, such as there is no need to display the search results corresponding to the query information in the form of cards.

[0128] It is understandable that the search extension pop-up layer, the focus display area, and the search extension tag in the tab bar in the first result page always exist, but whether the search extension pop-up layer, the focus display area, and the search extension tag are displayed in the first result page is conditional; the condition is whether the current query information can return the focus data. When the current query information can return the focus data, it can be displayed in the first result page; when the current query information cannot return the focus data, there is no need to display it in the first result page.

[0129] In an embodiment of the present application, the search expansion floating layer, the focus display area, and the search expansion tag in the tab bar in the first result page can be used as three entrances to view the focus topic directory associated with the current query information; each entrance can correspond to two states: a displayable state and a hidden state; the entrance in the displayable state can be displayed in the current page (for example, the first result page), and the entrance in the hidden state does not need to be displayed in the current page. It can be understood that, assuming that the current query information can return focus data, then in the first result page corresponding to the query information, only any one of the above three entrances can be displayed according to specific needs, or any two of the above entrances can be displayed at the same time, or the above three entrances can be displayed at the same time, and the present application does not limit this.

[0130] For example, after entering the first result page of the browser, the search extension floating layer can be displayed in the first result page, and the search extension tag in the tab bar can be hidden (the focus display area has not been exposed at this time); when the object slides the first result page until the focus display area is exposed, the focus display area is displayed in the first result page at this time, and the search extension floating layer and the search extension tag can be hidden; if the object continues to slide the first result page, the focus display area exits the current device display interface, then the search extension tag can be displayed in the first result page, and the search extension floating layer continues to be hidden. If you need to switch to the channel page corresponding to the search extension tag, you can display the search extension tag in the tab bar, and the search extension floating layer and the focus display area do not need to be displayed at this time. The above description is an example of displaying only one entrance in the first result page.

[0131] Optionally, assuming that the current query information can return the focus data, the search extension tag can be always displayed in the tab bar of the first result page, and the search extension floating layer and the focus display area can be displayed alternately; that is, two entrances can be displayed simultaneously in the first result page (for example, the search extension tag and the search extension floating layer can be displayed simultaneously, or the search extension tag and the focus display area can be displayed simultaneously). Optionally, when the focus display area in the first result page is exposed, the three entrances of the search extension floating layer, the focus display area, and the search extension tag can be displayed simultaneously in the first result page.

[0132] Step S103, in response to the detail viewing operation on the search extension floating layer, a directory of concerned topics associated with the query information is displayed in the channel page corresponding to the search extension tag.

[0133] Among them, the search extension floating layer may include an entry area and a detail viewing control. The entry area can be used to carousel recommended entries associated with the query information, and the detail viewing control can be used to view the directory of follow topics associated with the query information; the triggering operation of the detail viewing control by the target object can be called a detail viewing operation.

[0134] Specifically, when the object performs a trigger operation (e.g., a click operation) on the detail viewing control in the search expansion floating layer, the computer device can respond to the trigger operation (detail viewing operation) on the detail viewing control in the search expansion floating layer, switch from the current first result page to the channel page corresponding to the search expansion tag, and display the directory of topics of interest associated with the query information in the channel page corresponding to the search expansion tag; the directory of topics of interest can be generated based on multiple topics of interest associated with the query information and the recommended terms corresponding to each topic of interest. In other words, multiple topics of interest associated with the query information and the recommended terms corresponding to each topic of interest can be presented to the object in a structured recommendation manner.

[0135] In a feasible implementation, the computer device responds to the detail viewing operation on the search expansion floating layer and can obtain multiple topics of interest associated with the query information and the recommended terms corresponding to each topic of interest, and each topic of interest can correspond to one or more recommended terms. The multiple topics of interest associated with the query information are determined as first-level titles, and the recommended terms corresponding to each topic of interest are determined as second-level titles; based on the first-level titles and the second-level titles, a directory of topics of interest associated with the query information can be constructed, and the directory of topics of interest can be displayed in the channel page corresponding to the search expansion tag.

[0136] See also Figure 3 , Figure 3 This is a schematic diagram of an interface of a floating layer entrance in a result page provided in an embodiment of the present application. Figure 1 . Figure 3 The terminal device 20a shown can be a computer device used by the target object, and a client or browser with a search function can be run in the terminal device 20a (for ease of understanding, the browser is taken as an example below). When the target object starts the browser in the terminal device 20a, the search page 20b of the browser can be displayed in the terminal device 20a, and the search page 20b can include a search area 20c, and the search area 20c can include an audio entry 20d, a picture entry 20f, and a search control 20e. Among them, the audio entry 20d can be used to input the audio to be queried. If the target object inputs the audio to be queried through the audio entry 20d, then the audio input by the target object can be searched; the picture entry 20f can be used to input the picture to be queried. If the target object inputs the picture to be queried through the picture entry 20f, then the picture input by the target object can be searched.

[0137] like Figure 3 As shown, assuming that the target object enters the text content "the reason for the conflict between Xiao A and Xiao B" in the search area 20c, the text content displayed in the search area 20c can be used as the query information 20g. When the target object performs a trigger operation on the search control 20e in the search area 20c, the terminal device 20a can respond to the trigger operation on the search control 20e and display the first result page 20h corresponding to the query information 20g. Among them, the search area 20c can still be displayed in the first result page 20h, and the current query information 20g (such as "the reason for the conflict between Xiao A and Xiao B") can be displayed in the search area 20c. The first result page 20h may also display a tab bar 20i, which may include multiple different tabs (tabs), and different tabs may correspond to different channel pages. For example, the tab bar 20i may include, but is not limited to, "All", "Ultra Wide Angle", "Ask", "Encyclopedia", "Mini Program" and other tabs; the "Ultra Wide Angle" involved in the embodiment of the present application may be understood as the name of the above-mentioned search extension tag, and optionally, it may also be set to other names, which is not limited in the present application. When the first result page 20h is displayed in the browser of the terminal device 20a, the first result page 20h is displayed by default as the channel page corresponding to the "All" tab. At this time, the "All" tab in the tab bar 20i may be marked, such as underlining the "All" tab, or making the font of the "All" tab bold, etc., which is not limited in the present application.

[0138] like Figure 3As shown, the search results corresponding to the query information 20g can be displayed in the area 20j of the first result page 20h in the form of a card. It is understandable that due to the limited size of area 20j, the search results corresponding to the query information 20g may not be fully displayed in area 20j, then an "expand all" control can be displayed in area 20j; by performing a trigger operation on the "expand all" control, the complete search results can be viewed. If the search results corresponding to the query information 20g can be fully displayed in area 20j, then there is no need to display the "expand all" control in area 20j. In area 20k of the first result page 20h, some search terms associated with the query information 20g can be displayed. The search terms in area 20k may refer to multiple search terms with the highest historical search frequency among all the search terms associated with the query information 20g. As shown Figure 3 As shown, the first result page 20h may also display a search expansion floating layer 20m (the search expansion floating layer 20m may also be named as an "ultra-wide-angle" floating layer), and the search expansion floating layer 20m may include an "all" control 20n (details view control).

[0139] When the target object performs a trigger operation (details viewing operation) on the "All" control 20n in the search expansion floating layer 20m, the terminal device 20a can respond to the trigger operation on the "All" control 20n, and the tab bar 20i switches from the "All" tab to the "Ultra Wide Angle" tab, and displays the channel page 20p corresponding to the "Ultra Wide Angle" tab. The channel page 20p corresponding to the "Ultra Wide Angle" tab may include a top area 20q and a directory area 20s. Among them, the top area 20q here can be called a top bar, which can refer to a strip area located at the top of the channel page 20p corresponding to the "Ultra Wide Angle" tab; the top area 20q may include a search area 20c, a tab bar 20i, and a result return control 20r (i.e., used to return the first result page 20h corresponding to the query information 20g). The directory area 20s can display a directory of focus topics associated with the query information 20g, and the first-level titles in the focus topic directory can be the focus topics associated with the query information 20g (for example, focus topic V1, focus topic V2, focus topic V3, etc.); the second-level titles in the focus topic directory can be recommended terms under each focus topic (for example, recommended term c1, recommended term c2... recommended term c6, etc.).

[0140] The target object can slide the channel page 20p corresponding to the "Ultra Wide Angle" tag to view the complete directory of topics of interest associated with the query information 20g. It is understandable that when the target object slides the channel page 20p corresponding to the "Ultra Wide Angle" tag, the top area 20q (top bar) in the channel page 20p corresponding to the "Ultra Wide Angle" tag can be switched to a button ceiling, and the result return control 20r is displayed in the channel page 20p corresponding to the "Ultra Wide Angle" tag in a ceiling manner. The button ceiling here can be understood as the result return control 20r is fixedly displayed at the top position of the channel page 20p corresponding to the "Ultra Wide Angle" tag, and the display position will not be updated with the sliding of the page.

[0141] In a feasible implementation, for the search extension floating layer displayed in the first result page, the recommended terms of the focus topics associated with the query information can be displayed in the term area of ​​the search extension floating layer. In response to a trigger operation on the term area, the recommended term displayed in the term area is determined as a target term, and a second result page corresponding to the target term is displayed. In the second result page, the search results and target viewing controls corresponding to the target term can be displayed; the target viewing control can be used to view the focus topic directory in the channel page corresponding to the search extension tag. In the second result page, the search extension tag in the tab bar is in a hidden state, that is, there is no need to display the search extension tag in the tab bar of the second result page.

[0142] See also Figure 4 , Figure 4 This is a schematic diagram of an interface of a floating layer entrance in a result page provided in an embodiment of the present application. Figure 2 . Figure 4 The terminal device 30a shown may be a computer device used by the target object. For the operation related to displaying the first result page 30b corresponding to the query information 30d in the browser running on the terminal device 30a and the description related to the display content in the first result page 30b, please refer to Figure 3 The description of the first result page 20h in the corresponding embodiment will not be repeated here.

[0143] Since the query information 30d in the search area 30c (such as "the reason for the conflict between Xiao A and Xiao B") can return the focus point data, the "ultra-wide angle" label (i.e., the above-mentioned search extension label) can be displayed in the tab bar 30e of the first result page 30b, and the search extension floating layer 30f can be displayed in the first result page 30b. The search extension floating layer 30f may include a term area 30g and an "all" control (details viewing operation); the term area 30g can be used to rotate the recommended terms associated with the current query information 30d.

[0144] It is understandable that the query information 30d can be associated with multiple topics of interest, and each topic of interest can correspond to one or more recommended terms; that is, the query information 30d can be associated with multiple recommended terms. In the term area 30g of the search expansion floating layer 30f, all recommended terms associated with the query information 30d can be played in turn according to the pre-set playback rules. The pre-set playback rules may include but are not limited to: the speed of the recommended terms, the order of the recommended terms, the number of recommended terms updated each time, the scrolling direction of the recommended terms, etc.

[0145] For example, the recommended terms in the term area 30g of the search expansion floating layer 30f can be updated every 2 seconds, or can be updated every 4 seconds. The present application does not limit the carousel speed of the recommended terms in the term area 30g. According to the size of the term area 30g, it can be determined whether one recommended term is updated at a time, or two recommended terms are updated at a time in the term area 30g. The present application does not limit this. The scrolling direction of the recommended terms can be the horizontal direction of the terminal screen or the vertical direction. The present application does not limit this. The carousel order of the recommended terms can be sorted according to the first letter of the first character of each recommended term, or can be sorted in order from large to small according to the estimated value of the feedback indicator corresponding to each recommended term, or all recommended terms can be randomly sorted, etc. The present application does not limit the carousel order of the recommended terms.

[0146] Among them, the estimated indicator estimated value may refer to the estimated value obtained by analyzing the historical interaction data of each recommended term and the historical interaction data of the target object using an algorithm (for example, a deep learning algorithm). The estimated indicator evaluation value may include an estimated click-through rate (ctr), a web page conversion rate, etc.; the estimated click-through rate may indicate the estimated ratio from the exposure of the recommended term to the click of the recommended term, that is, the estimated click volume of the recommended term divided by the estimated display volume of the recommended term; the web page conversion rate may refer to the percentage of converted visitors to the total visitors of the web page. The historical interaction data of each recommended term may refer to a series of interaction data generated after the recommended term is exposed, such as the search, click, browse, favorite, share and other operation data of different objects on the recommended term. The historical interaction data of the target object may refer to the data generated when the target object interacted with the exposed recommended terms in the past, such as the terms that the target object has searched, clicked, browsed, favorited, shared, etc. in the past period of time (for example, the past month, the past 10 days, the past week, etc., which are not limited in this application).

[0147] It should be noted that when the recommended terms associated with the query information 30d are played in turn in the term area 30g, the arrow at the end of the recommended term being scrolled in the term area 30g can be temporarily hidden, indicating that the search results of the recommended term currently being scrolled cannot be viewed temporarily; when the scrolling stops, an arrow can be displayed at the end of the current recommended term in the term area 30g, indicating that the recommended term currently displayed in the term area 30g can be clicked to view the search results corresponding to the recommended term.

[0148] like Figure 4 As shown, when the target object performs a trigger operation on the recommended term c5 currently displayed in the term area 30g, the terminal device 30a can respond to the trigger operation on the recommended term c5 in the term area 30g. At this time, the recommended term c5 can be called the target term; the recommended term c5 is searched using the search function of the browser to obtain the search result corresponding to the recommended term c5. The search result corresponding to the recommended term c5 can be in the form of a card and displayed in the second result page 30h corresponding to the recommended term c5; optionally, the second result page 30h can also display the data source of the search result corresponding to the recommended term c5. At this time, the second result page 30h can be called the term landing page corresponding to the recommended term c5, and can also be called the secondary page of the first result page 30b; the display content in the second result page 30h is the display content corresponding to the "All" label in the tab bar 30e. The recommended term c5 may be displayed in the search area 30c of the second result page 30h. During the search for the recommended term c5, it is not necessary to retrieve the focus topics and recommended terms associated with the recommended term c5.

[0149] The second result page 30h may include a top area (top bar), and the top area of ​​the second result page 30h may include a search area 30c, a tab bar 30e, and a target viewing control 30i. The target viewing control 30i can be used to view the directory of topics of interest in the channel page corresponding to the "ultra wide angle" tag. The "ultra wide angle" tag (search extension tag) is not displayed in the tab bar 30e, and there is no need to display the search extension floating layer and the focus display area in the second result page 30h. In other words, the result page (for example, the second result page 30h) displayed after clicking the recommended term in the first result page 30b corresponding to the query information 30d does not need to display the "ultra wide angle" tag, the search extension floating layer, and the focus display area. The recommended term clicked above can be a recommended term in the directory of topics of interest displayed on the channel page corresponding to the "ultra wide angle" tag, or can be a recommended term displayed in the term area of ​​the search extension floating layer, or can be a recommended term displayed in the focus display area, etc.

[0150] It should be noted that the style of the interface schematic diagram involved in the embodiment of the present application is only an example, and the present application does not limit the specific style of the page (for example, the first result page, the second result page, the channel page corresponding to the search extension tag, etc.). For example, the top area of ​​the second result page can have different visual styles, such as but not limited to no top, dark top, light top, etc.

[0151] In a feasible implementation, when the object performs a trigger operation on the target viewing control in the second result page, the object can respond to the trigger operation on the target viewing control in the second result page by searching for an extended tag (e.g., Figure 3 The channel page corresponding to the "Ultra Wide Angle" label shown in the figure displays a directory of topics of interest associated with the query information. The target entry in the directory of topics of interest is highlighted, and in the directory of topics of interest, an extended entry is added to the topic of interest to which the target entry belongs. Among them, the highlighting of the target entry may refer to a display method that is distinguished from unviewed recommended entries, that is, the viewed recommended entries in the directory of topics of interest (for example, the target entry) and the unviewed recommended entries are distinguished and displayed. The specific presentation form of the highlighting may be any of the following: distinguishing by different fonts, distinguishing by different colors, distinguishing by whether there is an underline, distinguishing by whether the text is bold, distinguishing by whether it is set to italics, distinguishing by whether text, graphics, etc. are added, etc. The present application does not limit the specific presentation form of the highlighting.

[0152] Optionally, in the channel page corresponding to the search extension tag, the process of adding an extended term to the topic of interest to which the target term belongs may include but is not limited to: displaying a term connection animation in the topic of interest directory according to the display position of the target term in the channel page corresponding to the search extension tag; the term connection animation is used to rotate the extended terms of the topic of interest to which the target term belongs. When the term connection animation ends, the extended terms included in the term connection animation are expanded and displayed in the topic of interest directory, and the display positions of multiple topics of interest in the topic of interest directory and the recommended terms of multiple topics of interest are adjusted.

[0153] exist Figure 4 Based on the corresponding embodiments, please refer to Figure 5 , Figure 5 is a schematic diagram of an interface of a topic directory provided by an embodiment of the present application. Figure 5As shown, when the target object of the terminal device 30a is used to perform a trigger operation on the target viewing control 30i in the second result page 30h, the terminal device 30a can respond to the trigger operation on the target viewing control 30i and switch from the second result page 30h corresponding to the recommended term c5 to the concerned topic directory corresponding to the original query information ("the reason for the conflict between Xiao A and Xiao B"); the concerned topic directory is displayed in the channel page 30j corresponding to the "super wide angle" label. In other words, it is possible to switch from the second result page 30h to the channel page 30j corresponding to the "super wide angle" label.

[0154] Among them, the channel page 30j corresponding to the "Ultra Wide Angle" label may include a top area 30k and a directory area 30n. The top area 30k may include a search area 30c (displaying the query information "Reason for the conflict between Little A and Little B"), a tab bar, and a result return control 30m. The directory area 30n may display a directory of topics of interest associated with the query information ("Reason for the conflict between Little A and Little B"). Since the recommended entry c5 (target entry) in the directory of topics of interest is a recommended entry that has been viewed, the recommended entry c5 in the directory of topics of interest may be highlighted in the channel page 30j corresponding to the "Ultra Wide Angle" label, such as Figure 5 The graying out process shown here is only an example of highlighting, and the present application does not limit the highlighting form of the viewed recommended terms.

[0155] Optionally, the target object can continue to select other recommended terms from the concerned topic catalog associated with the query information ("the reason for the conflict between Xiao A and Xiao B") to view, and enter the second result page corresponding to any recommended term selected by the target object to view the search results of the recommended term selected by the target object; each viewed recommended term in the concerned topic catalog can be highlighted in the same way. If the target object performs a refresh operation on the channel page 30j corresponding to the "ultra-wide angle" label, the highlight of the viewed recommended terms in the concerned topic catalog can be cleared (e.g., the grayed-out viewed recommended terms can be restored to their original style).

[0156] Furthermore, in the concerned topic catalog associated with the query information ("the reason for the conflict between Xiao A and Xiao B"), an entry connection animation can be displayed for the recommended entry c5, and the entry connection animation can be used to rotate the extended entries of the concerned topic (such as concerned topic V2) to which the recommended entry belongs. Figure 5As shown, the entry connection animation can be displayed in the area 30p of the channel page 30j corresponding to the "Ultra Wide Angle" label; the extended entry displayed in the entry connection animation can refer to other recommended entries except the recommended entry c4 and the recommended entry c5 corresponding to the focus topic V2, and the extended entry in the entry connection animation and the recommended entry c5 belong to the focus topic V2. For example, assuming that the focus topic V2 associated with the query information ("the reason for the conflict between Xiao A and Xiao B") corresponds to 5 recommended entries, if they are sorted in descending order according to the estimated value of the feedback index, the order of the 5 recommended entries corresponding to the focus topic V2 can be recorded as: recommended entry c4, recommended entry c5, recommended entry c9, recommended entry c10 and recommended entry c11. The initial focus topic directory may include the first two recommended entries corresponding to the focus topic V2: recommended entry c4 and recommended entry c5; after the target object clicks to view the recommended entry c5, the recommended entry c9 and recommended entry c10 may be displayed alternately in the entry connection animation corresponding to the recommended entry c5. At this time, the recommended entry c9 and recommended entry c10 may be called the extended entries corresponding to the focus topic V2.

[0157] When the target object performs a trigger operation (for example, a click operation) on the entry connection animation being played, or when the entry connection animation ends, all the extended entries contained in the entry connection animation can be expanded and displayed in the directory of topics of interest displayed in the directory area 30n, and these extended entries can be added to the directory of topics of interest as secondary titles corresponding to the topic of interest V2; that is, one or more extended entries can be automatically loaded for the topic of interest V2 to which the viewed recommended entry c5 belongs. At the same time, the display positions of multiple topics of interest in the directory of topics of interest and the recommended entries of each topic of interest can be adjusted according to the number of newly added extended entries. The number of extended entries added to the topic of interest each time can be one, two, or three. The present application does not limit the number of extended entries in each entry connection animation. These extended entries are recommended entries that have not been viewed yet, and need to be distinguished from the viewed extended entry c5 and displayed.

[0158] In a feasible implementation, assuming that the search extension tag in the tab bar is in a displayable state, that is, when the search extension tag is displayed in the tab bar of the first result page, the object can directly perform a trigger operation on the search extension tag in the tab bar, and the computer device at this time can respond to the trigger operation on the search extension tag in the first result page, and display the channel page corresponding to the search extension tag; in the channel page corresponding to the search extension tag, display the directory of topics of interest associated with the query information and the result return control. The directory of topics of interest includes multiple topics of interest associated with the query information, and recommended terms corresponding to the multiple topics of interest; the result return control is used to return from the channel page corresponding to the search extension page to the first result page.

[0159] based on Figure 3 Based on the corresponding embodiments, please refer to Figure 6 , Figure 6 This is a schematic diagram of an interface of a tab entry in a tab bar of a result page provided in an embodiment of the present application. Figure 6 The current interface shown represents the first result page 20h corresponding to the query information ("the reason for the conflict between Xiao A and Xiao B"); the query information ("the reason for the conflict between Xiao A and Xiao B") can be displayed in the search area 20c of the first result page 20h, and the first result page 20h can display a tab bar 20i, an area 20j (for displaying the search results corresponding to the query information ("the reason for the conflict between Xiao A and Xiao B")), an area 20k (for displaying some search terms associated with the query information 20g), and a search expansion floating layer, etc. When the target object performs a trigger operation on the "ultra wide angle" tag (search expansion tag) in the tab bar 20i, the terminal device 20a can respond to the trigger operation on the "ultra wide angle" tag, switch from the "all" tag to the "ultra wide angle" tag in the tab bar 20i, and display the channel page 20p corresponding to the "ultra wide angle" tag. The channel page 20p corresponding to the "Ultra Wide Angle" tag may include a top area 20q and a directory area 20s; the top area 20q may include a search area 20c, a tab bar 20i, and a result return control 20r, and the directory area 20s may display a directory of topics of interest associated with the query information 20g.

[0160] Please note that, see Figure 7 , Figure 7 1 is a schematic diagram of an interface in which a directory of topics of interest fails to load, provided by an embodiment of the present application. Figure 7As shown, after the terminal device 20a responds to the trigger operation for the "ultra wide angle" tag, the channel page 20p corresponding to the "ultra wide angle" tag is displayed. If the directory of the concerned topics in the channel page 20p corresponding to the "ultra wide angle" tag fails to load, the top area 20q and the directory area 20s can still be displayed in the channel page 20p corresponding to the "ultra wide angle" tag. The display content in the top area 20q is consistent with the display content when the directory of the concerned topics is successfully loaded, and the directory area 20s can display the prompt message "Sorry, the current page fails to load" and the "Refresh page" control. By performing a trigger operation on the "Refresh page" control, the channel page 20p corresponding to the "ultra wide angle" tag can be refreshed until the directory of the concerned topics is successfully loaded in the directory area 20s. If the "Refresh page" control is triggered multiple times and the directory of the concerned topics still cannot be successfully loaded, the result return control 20r in the top area 20q can be triggered to return to the first result page corresponding to the query information ("the reason for the conflict between Xiao A and Xiao B").

[0161] In a feasible implementation, if a search is initiated for the target object in the search area of ​​the first result page corresponding to the query information, or the search area of ​​the second result page corresponding to the target term, or the search area of ​​the channel page corresponding to the search extension tag, then a new search can be initiated by entering the "All" tag, which is consistent with the logic of the query information entered in the search page.

[0162] See also Figure 8 , Figure 8 is a schematic diagram of an interface of a search process provided by an embodiment of the present application. The following describes the search process in the second result page corresponding to the target term, such as Figure 8 As shown, the terminal device 40a may be a computer device used by the target object. When the target object performs a trigger operation on the recommended term c5 associated with the query information (which may be a recommended term displayed in the term area of ​​the search expansion floating layer, or a recommended term displayed in the focus topic directory, or a recommended term displayed in the focus point display area, etc.), a second result page 40b corresponding to the recommended term c5 may be displayed; for a description of the display content of the second result page 40b, see Figure 4 The relevant description of the second result page 30h in the corresponding embodiment will not be repeated here.

[0163] When the target object performs an input operation on the search area 40c in the second result page 40b, the terminal device 40a can respond to the input operation on the search area 40c in the second result page 40b, can clear the existing content in the search area 40c (i.e., the recommended term c5), and re-enter new query information in the search area 40c for search.

[0164] like Figure 8 As shown, when the target object enters query information q1 in the search area 40c and performs a trigger operation on the search control in the search area 40c (which can be called a search operation for query information q1), the terminal device 40a can use the search function of the browser to search for the query information q1 and obtain the search results corresponding to the query information q1. If the query information q1 is retrieved to have focus data, such as the focus topic associated with the query information q1 and the recommended terms corresponding to the focus topic, then the search results corresponding to the query information q1 can be displayed in the first result page 40e corresponding to the query information q1 in the form of a card, the "Ultra Wide Angle" label can be displayed in the tab bar 40d, and the search expansion floating layer 40f can be displayed in the first result page 40e.

[0165] Alternatively, if Figure 8 As shown, when the target object enters the query information q2 in the search area 40c and performs a trigger operation on the search control in the search area 40c (which can be called a search operation for the query information q2), the terminal device 40a can use the search function of the browser to search for the query information q2 and obtain the search results corresponding to the query information q2. If the focus point data of the query information q2 is not retrieved, such as the focus topic associated with the query information q2 and the recommended terms corresponding to the focus topic, then the search results corresponding to the query information q2 can be displayed in the first result page 40h corresponding to the query information q2, such as Figure 8 In the area 40g shown; there is no need to display the "Ultra Wide Angle" label in the tab bar 40d, and there is no need to display the search expansion floating layer in the first result page 40h at this time.

[0166] It is understandable that the search operation initiated from the search area of ​​any page of the browser can adopt the same search logic as the aforementioned query information ("the reason for the conflict between Xiao A and Xiao B"), which will not be repeated here. Optionally, in the search process initiated from the recommended term associated with the current query information, the result page corresponding to the recommended term (for example, the second result page) will no longer display the search extension label, search extension floating layer, and focus display area, etc., so that the search traffic of the original query information can be improved while ensuring the traffic of its related searches.

[0167] In a feasible implementation, when the object slides the first result page corresponding to the query information, the computer device can respond to the sliding operation on the first result page and update the display content in the first result page according to the sliding speed indicated by the sliding operation; when the display content in the first result page includes a focus display area, the search expansion floating layer is hidden and displayed in the first result page; wherein the focus display area is used to display multiple focus topics associated with the query information, and recommended terms corresponding to the multiple focus topics. In other words, the search expansion floating layer and the focus display area in the first result page corresponding to the query information are not displayed at the same time, and when the focus display area in the first result page is exposed, the search expansion floating layer in the first result page can be hidden.

[0168] It is understandable that the focus display area of ​​the first result page may also include a detail viewing control. When the object triggers a detail viewing operation in the focus display area, the computer device may respond to the triggering operation of the detail viewing control in the focus display area, display the channel page corresponding to the search extension tag in the tab bar of the first result page, and display the focus topic directory associated with the query information in the channel page corresponding to the search extension tag; wherein the focus topic directory includes multiple focus topics included in the focus display area, and recommended terms corresponding to the multiple focus topics. If it is detected that there are viewed recommended terms among the recommended terms of the multiple focus topics associated with the query information, the viewed recommended terms and unviewed recommended terms can be distinguished and displayed in the focus display area.

[0169] See also Figure 9a and Figure 9b , Figure 9a This is a schematic diagram of an interface of an entrance to a focus display area in a result page provided in an embodiment of the present application. Figure 1 . Figure 9a The interface shown is the first result page 50b corresponding to the query information 50d ("the reason for the conflict between Little A and Little B"). The first result page 50b may include a search area 50c and a tab bar 50e. The search area 50c may display the query information 50d. If the query information 50d can return the focus point data, the "ultra-wide angle" label (search extension label) may be displayed in the tab bar 50e, and the bottom area of ​​the first result page 50b may be fixed to display the search extension floating layer 50f. The first result page 50b may also display the search results corresponding to the query information 50d, and the search results here may include the selected answers to the query information 50d and related information.

[0170] The target object (the object using the terminal device 50a) can update the display content in the first result page 50b by sliding (e.g., performing a sliding operation in the vertical direction of the terminal screen) the first result page 50b to view the information of interest to the target object. During the process of the target object sliding the first result page 50b, the bottom area of ​​the first result page 50b can continue to display the search expansion floating layer 5f until the focus display area 50g in the first result page 50b is exposed, and the search expansion floating layer 50f in the first result page 50b can be hidden. The focus display area 50g can be displayed in the first result page 50b in the form of a card.

[0171] Among them, the focus display area 50g can display some or all of the focus topics associated with the query information 50d, as well as the recommended terms corresponding to each focus topic. For example, the focus display area 50g can include the focus topics V1 and V2 associated with the query information 50d, and the focus topic V1 can include the recommended terms c1, c2, and c3, and the focus topic V2 can include the recommended terms c4 and c5. It can be understood that the target object can view all the focus topics associated with the query information 50d and the recommended terms under each focus topic by sliding (such as performing a sliding operation in the horizontal direction of the terminal screen) the focus display area.

[0172] like Figure 9a As shown, if the target object is interested in the recommended term c4 in the focus display area 50g, then the recommended term c4 can be clicked to switch to the second result page 50m corresponding to the recommended term c4. The recommended term c4 can be displayed in the search area 50c of the second result page 50m, and the "ultra wide angle" label (search extension label) is not displayed in the tab bar 50e of the second result page 50m, and there is no need to display the search extension floating layer. The target viewing control 50n can be displayed in the second result page 50m corresponding to the recommended term c4; by performing a trigger operation on the target viewing control 50n, the second result page 50m corresponding to the recommended term c4 can be switched to display the focus topic directory associated with the query information 50d (specifically, it can be displayed in the channel page 50i corresponding to the "ultra wide angle" label).

[0173] Alternatively, if Figure 9aAs shown, a "View All" control 50h (which may be referred to as a detail viewing control) may also be displayed in the focus display area 50g of the first result page 50b. If the target object wants to view the directory of topics of interest associated with the query information 50d, the "View All" control 50h in the focus display area 50g may be triggered, and the tab bar 50e in the first result page 50b switches from the "All" tab to the "Ultra Wide Angle" tab, and displays the channel page 50i corresponding to the "Ultra Wide Angle" tab. In the channel page 50i corresponding to the "Ultra Wide Angle" tab, a result return control 50j and a directory of topics of interest associated with the query information 50d may be displayed; the directory of topics of interest may be displayed in the directory area 50k, and the result return control 50j may be used to return the first result page 50b corresponding to the query information 50d ("the reason for the conflict between Little A and Little B").

[0174] It is understandable that the target object can click the recommended term c4 in the directory area 50k to switch to the second result page 50m corresponding to the recommended term c4. The target object can also click the target viewing control 50n in the second result page 50m to return to the channel page 50i corresponding to the "Ultra Wide Angle" label and distinguish and display the recommended term c4 in the focus topic directory from other unviewed recommended terms.

[0175] See also Figure 9b , Figure 9b This is a schematic diagram of an interface of an entrance to a focus display area in a result page provided in an embodiment of the present application. Figure 2 . When the target object triggers the result return control 50j in the channel page 50i corresponding to the "Ultra Wide Angle" label, the first result page 50b corresponding to the query information 50d can be returned, and the tab bar 50e switches from the "Ultra Wide Angle" label to the "All" label. If it is detected that the recommended terms c2 and c4 are both viewed recommended terms, the viewed recommended terms c2 and c4 can be highlighted (e.g., grayed out) in the focus display area 50g of the first result page 50b.

[0176] In an embodiment of the present application, when an object enters query information in a search box, a search operation for the query information can be responded to, and a first result page corresponding to the query information can be displayed, and the search results and search expansion floating layer corresponding to the query information can be displayed in the first result page. The search expansion floating layer is independently displayed on the first result page, and the tab bar of the first result page can include a search expansion tag; through the search expansion floating layer and the search expansion tag, you can jump to the channel page corresponding to the search expansion tag, and display the concerned topic directory associated with the query information in the channel page corresponding to the search expansion tag. Since the concerned topic directory contains multiple concerned topics associated with the query information, as well as the recommended terms under each concerned topic, and the concerned topic directory is displayed in the channel page corresponding to the search expansion tag; by establishing a clear "outline directory" perception of all recommended terms associated with the query information, that is, the concerned topic directory, it is beneficial for the object to click and view the recommended terms of interest in the topic directory in turn, which can improve the click-through rate of the concerned topic directory, and thus improve the search traffic. By organizing the recommended terms associated with the query information into a concerned topic directory, the efficiency of meeting the needs of the object can be improved, the user satisfaction of the browser or client can be improved, and the user retention rate can be improved.

[0177] It is understandable that the following describes in detail the multiple topics of interest associated with the query information and the method for determining the recommended terms corresponding to each topic of interest.

[0178] See also Fig.10 , Fig.10 This is a flow diagram of a data processing method provided in an embodiment of the present application. Figure 2 It can be understood that the data processing method can be executed by a computer device, which can be a terminal device, such as Figure 1 Any terminal device in the terminal cluster shown, or a server, such as Figure 1 The server 10d shown in the figure is not limited in this application. Fig.10 As shown, the method may include the following steps S201 to S210:

[0179] Step S201, obtaining a historical query statement and M historical candidate terms corresponding to the historical query statement.

[0180] The historical query statement may refer to the historical query statement selected from the historical search records recorded in the log file; the number of historical query statements may be multiple, and each historical query statement may be a word, phrase or sentence described in natural language. The historical candidate terms may refer to the recommended terms that meet the preset conditions selected from all the recommended terms of each historical query statement; the preset conditions here may include but are not limited to: taking the recommended terms whose page views (PV) are greater than or equal to the page views threshold as the historical candidate terms.

[0181] Specifically, the computer device can mine historical query statements and historical candidate terms at regular intervals, such as from the log files of the browser or client, and mine the query statements of the object at regular intervals. These query statements mined from the log files can be called historical query statements; the period of time here can be one hour, two hours, 12 hours, one day, etc., and this application does not limit this. Furthermore, the recommended terms corresponding to each historical query statement can be obtained, and the page views corresponding to each recommended term can be counted; the recommended terms with page views greater than or equal to the page views threshold (the value can be pre-set according to the actual application scenario, and this application does not limit this) are determined as the historical candidate terms corresponding to the historical query statement. Each historical query statement can correspond to one or more historical candidate terms. For ease of understanding, the number of historical candidate terms corresponding to a historical query statement can be recorded as M, where M is a positive integer.

[0182] It is understandable that if the page views of all the recommended terms corresponding to the historical query statement are less than the page view threshold, the historical query statement and the recommended terms corresponding to the historical query statement may be discarded.

[0183] Step S202: input the historical query statement into the language model, generate a first branch topic corresponding to the historical query statement through the language model, and generate a first branch term corresponding to the first branch topic through the language model.

[0184] In an embodiment of the present application, a language model can be used to sequentially generate the focus topics corresponding to the historical query statements and the specific recommended terms for each focus topic based on the above historical query statements. Among them, the focus topics involved in the present application may refer to topics that are widely concerned, discussed, and have a high influence within a specific period or range; or they can be used to represent the reasons for recommending specific recommended terms, such as extracting the commonalities of specific recommended terms based on semantic methods (for example, text semantic vectors, keywords, etc.), object interaction data (for example, exposure click data, etc.), etc., and then the language model can be used to summarize and refine the extracted commonalities to generate the recommendation reasons corresponding to the historical query information; the recommendation reasons can also be used as the above-mentioned focus topics.

[0185] For ease of understanding, the embodiment of the present application may refer to the topic of interest generated by the language model as the first branch topic, and the recommended entry generated by the language model as the first branch entry. The language model involved in the embodiment of the present application may be a large language model, which may include but is not limited to: GPT4 (Generative Pre-trained Transformer 4, the 4th generation generative pre-trained language model based on the Transformer structure), Qwen1.5-32B-Chat (an open source large language model), llama3-70B (a general language model), Skywork-13B (an open source large language model) or any other large language model. For example, the first branch topic corresponding to the historical query statement and the first branch entry corresponding to the first branch topic may be obtained by asking questions to the language model; the question form may be set according to the specific application scenario, or may be adjusted in real time according to the specific content of the historical query statement. The present application does not limit the form of the question statement.

[0186] In a feasible implementation, the historical query statement can be input into the language model, and the historical query statement can be converted into a historical embedded representation through the embedding layer in the language model; the historical embedded representation can be encoded through the encoding component in the language model to obtain the historical encoding features corresponding to the historical query statement; the historical encoding features can be transformed through the fully connected layer in the language model to generate the first branch topic corresponding to the historical query statement.

[0187] Specifically, the historical query statements can be segmented to obtain multiple unit characters (the unit characters here can be single words or single words in the historical query statements); each unit character is converted into a vector to obtain the word vector representation corresponding to each unit character. For example, one-hot (unique hot code) can be used to represent each unit character as a word vector. At this time, the word vector representation is a vector containing only one 1 and the rest are 0. Among them, the dictionary (also known as a character bag) can include a series of common unit characters. According to the position of each common unit character in the dictionary, the one-hot code corresponding to each common unit character can be obtained. The word vector representations (such as one-hot codes) corresponding to each unit character of the historical query statement are spliced ​​to obtain the text vector representation corresponding to the historical query statement; the text vector representation corresponding to the historical query statement is converted through the embedding layer in the language model to obtain the historical embedding representation corresponding to the historical query statement. Among them, the embedding layer in the language model can be a word vector model, which may include but is not limited to: word2vec (a word vector model that obtains a vectorized representation of a word through the context of the word), glove (a word vector using a co-occurrence matrix), ELMo (Embeddings from Language Models), BERT (Bidirectional Encoder Representations from Transformers), latent semantic analysis model (Latent Semantic Analysis, LSA), latent Dirichlet Allocation model (Latent Dirichlet Allocation, LDA). The embodiment of the present application does not limit the word vector model used.

[0188] Through the encoding component in the language model, the historical embedding representation is feature extracted to obtain the historical encoding feature. Among them, the encoding component in the language model can be used to convert the input historical embedding representation into a fixed-length context vector, and the context vector at this time can be called a historical encoding feature; the encoding component can be any one of the structures such as transformer (a deep learning model using an attention mechanism), residual network, recurrent neural network (RNN), convolutional neural network (CNN), gated network, etc., or any combination or deformation of two or more network structures. This application does not limit the structure of the encoding component in the language model. Furthermore, the historical encoding feature can be feature transformed through the fully connected layer in the language model to generate the first branch topic corresponding to the historical query statement. It should be understood that the first branch topic can be input into the language model, and the same processing process as the historical query statement is used to generate the first branch entry corresponding to the first branch topic, which will not be repeated here.

[0189] For example, in a related search scenario, for each historical query statement, a language model can be used to generate 2-3 first branch topics that may be of interest to the object (the object that inputs the historical query statement); based on each first branch topic, two or more first branch terms can be generated. The number of first branch topics and the number of first branch terms corresponding to each historical query statement can be pre-set, and this application does not limit this.

[0190] Step S203 , clustering the M historical candidate terms corresponding to the historical query sentence to obtain N candidate term groups, and generating a second branch topic for each candidate term group.

[0191] Specifically, for each historical query statement, M historical candidate terms can be recalled, and then the language model can be used to cluster the M historical candidate terms corresponding to each historical query statement to obtain N candidate term groups, and a second branch topic is generated for each candidate term group. Among them, a candidate term group can correspond to a category; for example, the M historical candidate terms corresponding to each historical query statement can be divided into N categories, N can be a positive integer less than or equal to M, N can be a pre-set fixed value, or it can be determined by continuously merging the M historical candidate terms during the clustering process. This application does not limit the specific value of N.

[0192] When the computer device performs clustering processing on the M historical candidate terms corresponding to each historical query statement, the clustering algorithm adopted may include but is not limited to: K-means (mean) clustering algorithm, mean shift clustering algorithm, DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm, expectation-maximization (EM) clustering algorithm using Gaussian Mixture Model (GMM), hierarchical clustering algorithm, large language model, etc. The embodiment of the present application does not limit the type of clustering algorithm adopted.

[0193] It is understandable that the M historical candidate terms corresponding to each historical query sentence can be subjected to the same clustering process. For ease of understanding, the clustering process of the M historical candidate terms corresponding to any historical query sentence is specifically described below using K-means as an example.

[0194] The M historical candidate terms corresponding to the historical query statements can be converted into historical text vectors. The calculation process of the historical text vectors can be involved in the calculation process of the aforementioned historical embedding representation, which will not be repeated here. The historical text vectors corresponding to the M historical candidate terms can be added to the text vector set, and the center vector t is selected from the text vector set. i ; i is a positive integer less than or equal to N, where N can be a pre-set number of clusters. The number of clusters set in the clustering algorithm may affect the final clustering effect of the historical text vectors. If the number of clusters is too large, the historical candidate terms corresponding to similar historical text vectors may be divided into different candidate term groups. If the number of clusters is too small, the historical candidate terms corresponding to dissimilar historical text vectors may be divided into the same candidate term group. In the embodiment of the present application, the number of clusters can be set according to the specific application scenario. For example, N can be a value in the range of 1-4. Of course, the number of clusters can also be determined by cross-validation and other methods. N historical text vectors can be randomly selected from the text vector set as the central vector, or N historical text vectors that are as far away from each other as possible can be selected from the text vector set as the initial central vector, which can be respectively recorded as central vectors t1,..., central vector t i , ..., center vector t N .

[0195] Each historical text vector and each central vector (such as the central vector t i) between the two historical text vectors; the embodiment of the present application can use the distance between the two historical text vectors to measure the third similarity between the two. The larger the distance, the smaller the third similarity, and the smaller the distance, the greater the third similarity. j With the center vector t i The third similarity between them is the largest, so the historical text vector G j Add to the center vector t i The pending cluster C i , that is, the historical text vector G j Divide into the cluster to be processed to which the nearest center vector belongs, and initially obtain N clusters to be processed, where j is a positive integer less than or equal to M. Each center vector can correspond to a cluster to be processed, and each cluster to be processed can include one or more historical text vectors. The number of historical text vectors contained in different clusters to be processed can be the same or different. Among them, the method used to calculate the third similarity between two historical text vectors may include but is not limited to: Euclidian distance, Manhattan distance, Minkowski distance, cosine similarity, etc.

[0196] According to the cluster C to be processed i The historical text vector contained in the update center vector t i , until the cluster to be processed C i The updated center vector t i and the center vector t before update i If the same, according to the cluster C to be processed i Get N candidate word groups. For example, each cluster to be processed (such as the cluster to be processed C i ) contains the mean of all historical text vectors, and updates the cluster C to be processed according to the mean. i The corresponding center vector is obtained to obtain a new center vector, and then the historical text vectors contained in each cluster to be processed are updated according to the distance between each historical text vector in the text vector set and each new center vector. The above process is repeated continuously. When the historical text vectors contained in each cluster to be processed no longer change, that is, the center vectors corresponding to each cluster to be processed are fixed, the N clusters to be processed at this time can be determined as the final clustering result of the text vector set. The historical candidate terms corresponding to the historical text vectors in a final cluster to be processed can form a candidate term group; thus, N candidate term groups can be obtained based on the N final clusters to be processed.

[0197] Furthermore, the historical candidate terms in each candidate term group can be input into the language model, and the second branch topic corresponding to each candidate term group can be generated through the language model. For example, the language model can be asked questions, and all the historical candidate terms in a candidate term group can be input each time to generate the second branch topic corresponding to each candidate term group; the second branch terms corresponding to the second branch topic can include the historical candidate terms in the candidate term group corresponding to the second branch topic.

[0198] Step S204, performing deduplication processing on the first branch terms and the second branch topics to obtain the concerned topics corresponding to the historical query statements.

[0199] Specifically, since the first branch topic and the second branch topic are generated by two-way logic, the first branch topic is a topic of interest generated for each historical query statement through a language model; the first branch topic can be considered to be obtained by generation. The second branch topic is a topic of interest generated by clustering M historical candidate terms and the final clustering result (N candidate term groups), and the second branch topic can be considered to be obtained by clustering.

[0200] The following takes any historical query statement (for example, historical query statement q0) as an example to describe the deduplication process between the first branch topic and the second branch topic corresponding to the historical query statement q0. The first branch topic and the second branch topic corresponding to the historical query statement q0 can be merged to obtain an initial topic set corresponding to the historical query statement q0. The topic semantic vectors corresponding to the first branch topic and the second branch topic in the initial topic set are obtained, and the first semantic similarity between any two first branch topics and the second branch topics in the initial topic set is calculated based on the topic semantic vectors corresponding to the first branch topic and the second branch topic in the initial topic set; wherein, the calculation method of the first semantic similarity can adopt the same calculation method as the aforementioned third similarity, which will not be repeated here.

[0201] When the first semantic similarity between the first branch topic V4 and the second branch topic V5 in the initial topic set is greater than or equal to the similarity threshold, any one of the first branch topic V4 and the second branch topic V5 can be retained and the other topic can be discarded. For example, the second branch topic V5 can be retained, and the first branch term corresponding to the first branch topic V4 and the historical candidate terms (which can be referred to as second branch terms) in the candidate term group corresponding to the second branch topic V5 can be combined into the candidate terms contained in the retained second branch topic V5. At this time, the second branch topic V5 can be called the focus topic corresponding to the historical query statement q0.

[0202] Through the above method, any one of the two topics whose first semantic similarity is greater than or equal to the similarity threshold can be filtered out from the first branch topic and the second branch topic corresponding to the historical query statement q0, and all the focus topics corresponding to the historical query statement q0 and the candidate terms contained in each focus topic can be obtained. Similarly, for each historical query statement, the above process can be used to obtain all the focus topics corresponding to each historical query statement and the candidate terms contained in each focus topic corresponding to each historical query statement.

[0203] Step S205, deduplication processing is performed on the M historical candidate terms and the first branch terms corresponding to the first branch topic to obtain recommended candidate terms for the topic of interest corresponding to the historical query statement.

[0204] Specifically, since the candidate terms contained in each focus topic corresponding to the historical query statement may include both the first branch terms obtained by the generation method and the second branch terms obtained by the clustering method; for example, the candidate terms under the second branch topic V5 (a focus topic corresponding to the historical query statement q0) corresponding to the above historical query statement q0 include the first branch terms corresponding to the first branch topic V4 (obtained by the generation method), and the second branch terms corresponding to the second branch topic V5 (obtained by the clustering method). Therefore, the candidate terms contained in each focus topic corresponding to the historical query statement can be deduplicated, which can be understood as deduplicating the aforementioned M historical candidate terms and the first branch terms corresponding to the first branch topic to obtain the recommended candidate terms for each focus topic corresponding to the historical query statement.

[0205] Among them, the deduplication process between the candidate terms included in each focus topic may include offline deduplication processing and online deduplication processing. The offline deduplication process may include but is not limited to: for any focus topic corresponding to the historical query statement, the second semantic similarity between the candidate terms included in a focus topic can be calculated, and the calculation method of the second semantic similarity can adopt the same calculation method as the aforementioned third similarity, which will not be repeated here. Based on the second semantic similarity, the offline deduplication process between the candidate terms included in a focus topic can refer to the deduplication process between the first branch topic and the second branch topic in step S204, which will not be repeated here.

[0206] It is understandable that, for multiple candidate terms whose second semantic similarity is greater than or equal to the similarity threshold, after offline deduplication processing, 1-2 candidate terms whose second semantic similarity is greater than or equal to the similarity threshold may still be retained. In other words, the offline deduplication process can be understood as a rough deduplication process for the candidate terms included in the topic of interest.

[0207] The online deduplication process may include but is not limited to: obtaining the estimated feedback index value (for example, estimated click rate) corresponding to the candidate terms obtained by the above offline deduplication process; for two candidate terms whose second semantic similarity is greater than or equal to the similarity threshold, the candidate term with a higher estimated feedback index value may be selected for retention, and the recommended candidate terms included in each focus topic corresponding to each historical query statement are obtained. At this time, each focus topic corresponding to each historical query statement, as well as the recommended candidate terms included in each focus topic, are all obtained after deduplication processing.

[0208] In a feasible implementation, the number of topics of attention after each historical query statement is deduplicated can be counted. If the number of topics at this time is less than the first number threshold, then there is no need to perform structured recommendation on the historical query information, that is, the historical query information, the attention topic corresponding to the historical query information, and the recommended candidate terms contained in the attention topic can be deleted. Similarly, the number of terms of the recommended candidate terms contained in the attention topic after deduplication can be counted. If the number of terms at this time is less than the first number threshold, then the attention topic and the recommended candidate terms corresponding to the attention topic can be deleted. If the number of terms at this time is greater than the second number threshold, then the recommended candidate terms contained in the attention topic after deduplication can be sorted according to the feedback index estimated value corresponding to each recommended candidate term (sorted in order from high to low according to the feedback index estimated value), and a fixed number of recommended candidate terms are retained. The fixed number here is equal to the above-mentioned second number threshold, the second number threshold is greater than the first number threshold, and the first number threshold and the second number threshold can be pre-set according to actual application requirements, and this application does not limit this.

[0209] Step S206: adding the historical query statements, the concerned topics corresponding to the historical query statements, and the recommended candidate terms of the concerned topics corresponding to the historical query statements to the index database.

[0210] Specifically, after the above processing, all remaining historical query statements, the focus topics corresponding to each historical query statement, and the recommended candidate terms of the focus topics corresponding to each historical query statement may be added to the index database.

[0211] Among them, the index database is a database for storing and managing index data, and the index is a data structure that sorts the values ​​of one or more columns in the database table, which can speed up data retrieval. The index database involved in this application may include an inverted index library and a semantic index library; the inverted index library refers to a database that uses historical query statements as index information, and each index information in the inverted index library can correspond to multiple focus topics and recommended candidate entries contained in each focus topic. The semantic index library is a database that uses the semantic feature vector corresponding to the historical query information as index information. Each index information in the semantic index library can correspond to a historical query information, and one historical query information corresponds to multiple focus topics and recommended candidate entries contained in each focus topic.

[0212] Step S207, obtaining query information, and searching the query information in the index database.

[0213] Specifically, when the object enters query information in the search area of ​​the search page, the computer device can obtain the query information, perform semantic analysis on the query information, and obtain the text semantic vector corresponding to the query information; if a query statement matching the query information is retrieved in the inverted index library, the query statement matching the query information in the inverted index library is determined as the target index information of the query information in the inverted index library; if a semantic index matching the text semantic vector is retrieved in the semantic index library, the semantic index matching the text semantic vector in the semantic index library is determined as the target index information of the query information in the semantic index library. In other words, for the query information entered by the object, it can be searched in the semantic index library and the inverted index library at the same time, and the index information matching the query information in the two index libraries is called the target index information. By adopting two retrieval methods, the retrieval coverage of the query information can be improved and the retrieval accuracy can be improved.

[0214] Optionally, any one of the inverted index library and the semantic index library may be selected for retrieval, and the index information in the index library that matches the query information is referred to as target index information, which is not limited in the present application.

[0215] Step S208: If target index information matching the query information is retrieved in the index database, multiple topics of interest corresponding to the target index information and multiple recommended candidate entries for each topic of interest corresponding to the target index information are obtained in the index database.

[0216] Step S209: Screen multiple recommended candidate terms for each topic of interest corresponding to the target index information to obtain a recommended term for each topic of interest corresponding to the target search information.

[0217] Specifically, if target index information matching the query information is not retrieved in the index database, it means that the query information cannot return the focus point data, and subsequent steps can be stopped.

[0218] If the target index information matching the query information is retrieved in the index database, it means that the query information can return the focus point data, and all the focus topics corresponding to the target index information and all the recommended candidate terms contained in each focus topic can be obtained. Further, the text semantic vector corresponding to the query information can be obtained, and multiple recommended candidate terms of each focus topic corresponding to the target index information can be added to the first candidate term set; the term semantic vector corresponding to each recommended candidate term in the first candidate term set is obtained, and the first similarity between the text semantic vector and the term semantic vector is obtained; in the first candidate term set, the recommended candidate terms whose first similarity is less than the first similarity threshold are deleted to obtain the second candidate term set. In other words, the first similarity between the query information and each recommended candidate term associated with it can be calculated by the text semantic vector corresponding to the query information and the term semantic vector corresponding to the recommended candidate term. When the first similarity between the query information and a recommended candidate term is less than the first similarity threshold, it means that the correlation between the recommended candidate term and the query information is not high, so the recommended candidate terms with low correlation with the query information can be deleted to obtain the second candidate term set.

[0219] Furthermore, the feedback index estimate corresponding to each recommended candidate term in the second candidate term set can be obtained, and the second similarity between each recommended candidate term in the second candidate term set can be obtained according to the term semantic vector; if in the second candidate term set, there is a second similarity between recommended candidate term a and recommended candidate term b that is greater than the second similarity threshold, and the feedback index estimate of recommended candidate term a is greater than the feedback index estimate of recommended candidate term b, then the recommended candidate term b is deleted from the second candidate term set to obtain a third candidate term set; among the recommended candidate terms included in the third candidate term set, the recommended terms for each topic of interest corresponding to the target retrieval information are determined. Among them, the determination process of the third candidate term set can be essentially considered as an online deduplication process between recommended candidate terms. For multiple recommended candidate terms whose second similarity is greater than the second similarity threshold, the recommended candidate term with the highest feedback index estimate (for example, estimated click rate) can be selected to obtain a third candidate term set.

[0220] The calculation methods of the first similarity and the second similarity can refer to the calculation method of the third similarity, which will not be repeated here. The first similarity threshold, the second similarity threshold and the similarity threshold can be set to the same value or different values, which is not limited in this application.

[0221] Step S210: Generate a catalog of topics of interest associated with the query information according to the multiple topics of interest corresponding to the target index information and the recommended terms of each topic of interest corresponding to the target search information.

[0222] Specifically, the estimated feedback index value of each focus topic corresponding to the target index information can be determined based on the estimated feedback index value corresponding to the recommended term of each focus topic; the multiple focus topics corresponding to the target index information are sorted according to the estimated feedback index value of each focus topic to obtain the sorted multiple focus topics; the recommended terms of each focus topic are sorted according to the estimated feedback index value corresponding to the recommended terms of each focus topic to obtain the sorted recommended terms corresponding to each focus topic; based on the sorted multiple focus topics and the sorted recommended terms corresponding to each focus topic, a directory of focus topics associated with the query information is generated.

[0223] See also Fig.11 , Fig.11 Schematic diagram of an architecture strategy for a related search scenario provided by an embodiment of the present application. Fig.11 As shown, the architectural strategy for relevant search scenarios may include near-line processes and online services; wherein, the near-line process may be between the online process (real-time processing) and the offline process (non-real-time processing), and may be used to process historical query statements within a recent period of time (for example, nearly one hour, etc.) and historical candidate terms corresponding to the historical query statements; the processing, interaction and feedback of the information involved in the online process all occur in a timely manner.

[0224] like Fig.11 As shown, near-line processes may include:

[0225] ① Multiple historical query statements and historical candidate terms corresponding to each historical query statement can be mined from the log; the number of historical candidate terms corresponding to a historical query statement can be recorded as M, and M can be an integer greater than 1. For example, historical query statements of an object can be mined from the log at regular intervals, and from all recommended terms corresponding to the historical query statements, recommended terms with page views greater than or equal to a page view threshold are filtered out as M historical candidate terms corresponding to the historical query statements.

[0226] ② The generation method and clustering method can be used to construct the focus topics and recommended candidate entries. The implementation logic of the generation method may include: directly generating the first branch topic and the first branch entries contained in the first branch topic for each historical query statement through the language model. The specific implementation process can be found in Fig.10The description of step S202 in the corresponding embodiment will not be repeated here. The implementation logic of the clustering method may include: clustering the M historical candidate terms corresponding to each historical query statement to obtain N candidate term groups, that is, dividing the M historical candidate terms into N categories; using the language model to generate a second branch topic for each candidate term group, the second branch term corresponding to the first branch topic may refer to all the historical candidate terms contained in the candidate term group corresponding to the second branch topic, and the specific implementation process can be seen in Fig.10 The description of step S203 in the corresponding embodiment will not be repeated here.

[0227] Furthermore, the first branch topic and the first branch term obtained by the generation method, as well as the second branch topic and the second branch term obtained by the clustering method can be merged and deduplicated to obtain the focus topic corresponding to each historical query statement and the recommended candidate terms contained in each focus topic. For the specific implementation process, please refer to Fig.10 The description of step S204 and step S205 in the corresponding embodiment will not be repeated here.

[0228] ③Write the topics of interest corresponding to the historical query statements and the recommended candidate terms contained in each topic of interest into the inverted index library and semantic index library in real time.

[0229] ④ Write production features into the feature service in real time. Production features may refer to relevant elements or indicators that can reflect the specific attributes, properties or status of the search system; production features may include but are not limited to: historical query statements input by the object and their related attribute information (such as word frequency, part of speech, etc.), historical interaction data (such as the object's search history, click operations, browsing time and other data), time information (specific time points and time periods of object searches), semantic features (historical query statements, topics of interest, and semantic features corresponding to recommended candidate entries), popularity features (the popularity of historical query statements), contextual information (contextual information formed based on the relevant operations of the object before and after this search) one or more. It can be understood that the production features corresponding to each semantic query statement can be written into the feature service in real time.

[0230] like Fig.11 As shown, when an object enters query information in a search page of a browser or client and initiates a search operation, the processing flow can be understood as an online service, which may include:

[0231] ① Inverse index / semantic recall: You can use inverse index retrieval and semantic retrieval to retrieve multiple topics of interest corresponding to the query information and the recommended candidate terms under each topic of interest. For the specific implementation process, please refer to Fig.10The description of step S207 and step S208 in the corresponding embodiment will not be repeated here.

[0232] ② Relevance filtering: The relevance between the query information and all the recommended candidate terms associated with the query information can be determined, and irrelevant recommended candidate terms can be filtered out. Among them, the remaining recommended candidate terms after filtering out irrelevant recommended candidate terms can form the second candidate term set. The specific implementation process of relevance filtering can be referred to in Fig.10 In step S209 of the corresponding embodiment, the description of obtaining the second candidate entry set is not repeated here.

[0233] ③ Sorting of feedback index estimated values ​​(e.g., estimated click-through rate): The estimated feedback index value corresponding to each recommended candidate term in the second candidate term set can be estimated through the sorting and scoring service and the production service, and the recommended candidate terms in the second candidate term set can be sorted in descending order of the estimated feedback index values. The estimated feedback index value of each topic of interest corresponding to the query information can be the average of the estimated feedback index values ​​corresponding to all the recommended candidate terms under it; similarly, multiple topics of interest corresponding to the query information can be sorted in descending order of the estimated feedback index values.

[0234] ④ Semantic deduplication: Based on the feature service, semantic deduplication can be performed on the recommended candidate terms in the second candidate term set (for example, online deduplication can be performed) to obtain the final multiple focus topics of the query information and the recommended terms corresponding to each focus topic. The specific implementation process can be found in Fig.10 The relevant description in step S209 of the corresponding embodiment will not be repeated here.

[0235] ⑤ Response package: Generate structured data (for example, a directory of topics of interest) based on the multiple topics of interest corresponding to the query information and the recommended terms corresponding to each topic of interest. The structured data can be displayed in the first result page corresponding to the query information.

[0236] In the embodiment of the present application, when the subject enters query information in the search page of the browser or client and initiates a search operation, an offline method can be used to generate a directory of topics of interest associated with the query information, such as Fig.11Relevant descriptions in the corresponding embodiments. Optionally, when the object enters query information in the search page of the browser or client and initiates a search operation, an online method can also be used to generate a directory of topics of interest associated with the query information. For example, after obtaining the query information, the two logics of generation and clustering can be used to generate online the topics corresponding to the query information and the recommended candidate terms corresponding to each topic; then, the topics corresponding to the query information and the recommended candidate terms corresponding to each topic can be subjected to relevance filtering, click-through rate estimation, deduplication and other operations to obtain multiple topics of interest corresponding to the query information and the recommended terms corresponding to each topic of interest, and generate structured data online, such as generating a directory of topics of interest.

[0237] In the embodiment of the present application, by establishing a clear "outline catalog" perception of all recommended terms associated with the query information, that is, a focus topic catalog, it is beneficial for the subject to click and view the recommended terms of interest in the topic catalog in turn, which can improve the click-through rate of the focus topic catalog and thus increase search traffic. By generating the focus topics corresponding to the historical query statements and the recommended terms contained in each focus topic through the generation method and clustering method, the recommended terms associated with each historical query statement can be organized and clustered, which can improve the coverage of the recommended terms.

[0238] It is understandable that in the specific implementation of this application, it may involve relevant information such as the user's login information, interest tags, historical search records, historical interaction data, etc. in the client or browser. When the above embodiments of this application are applied to specific products or technologies, it is necessary to obtain the permission or consent of the relevant institutions or departments, or the users themselves, and the collection, use and processing of relevant data need to comply with the relevant laws, regulations and standards of the relevant regions.

[0239] See also Fig.12 , Fig.12 Schematic diagram of a data processing device provided in an embodiment of the present application. Fig.12 As shown, the data processing device 1 may include: a page display module 101, a directory display module 102;

[0240] The page display module 101 is used to respond to the search operation for the query information and display the first result page corresponding to the query information;

[0241] The page display module 101 is further used to display the search results corresponding to the query information and the search expansion floating layer associated with the query information in the first result page; the search expansion floating layer is independently displayed on the first result page, and the tab bar of the first result page includes a search expansion tag;

[0242] The directory display module 102 is used to respond to the detail viewing operation on the search extension floating layer and display the concerned topic directory associated with the query information in the channel page corresponding to the search extension tag.

[0243] In one or more embodiments, the page display module 101 is further configured to perform the following steps:

[0244] In the entry area of ​​the search extension floating layer, the recommended entries of the topic of interest associated with the query information are displayed in a carousel;

[0245] In response to a trigger operation on the entry area, determining the recommended entry displayed in the entry area as a target entry, and displaying a second result page corresponding to the target entry;

[0246] In the second result page, the search results and directory viewing controls corresponding to the target term are displayed; the search extension tab in the tab bar of the second result page is in a hidden state.

[0247] In one or more embodiments, the page display module 101 is further configured to perform the following steps:

[0248] In response to a trigger operation on a directory viewing control in the second result page, a directory of topics of interest associated with the query information is displayed in a channel page corresponding to the search extension tag;

[0249] The target entry in the focused topic catalog is highlighted, and in the focused topic catalog, an extended entry is added to the focused topic to which the target entry belongs.

[0250] In one or more embodiments, the page display module 101 adds an extended term to the topic of interest to which the target term belongs in the topic of interest catalog, so as to perform the following steps:

[0251] According to the display position of the target term in the channel page corresponding to the search extension tag, the term connection animation is displayed in the focus topic directory; the term connection animation is used to rotate the extended terms of the focus topic to which the target term belongs;

[0252] When the entry connection animation is finished, the extended entries included in the entry connection animation are expanded and displayed in the focus topic directory, and the display positions of multiple focus topics in the focus topic directory and the recommended entries of the multiple focus topics are adjusted.

[0253] In one or more embodiments, the directory display module 102 responds to the detail viewing operation on the search expansion floating layer, and displays the directory of concerned topics associated with the query information in the channel page corresponding to the search expansion tag, for performing the following steps:

[0254] In response to a detail viewing operation on the search expansion floating layer, multiple topics of interest associated with the query information and recommended terms corresponding to each topic of interest are obtained;

[0255] Determine multiple topics of interest associated with the query information as first-level titles, and determine recommended terms corresponding to each topic of interest as second-level titles;

[0256] According to the primary title and the secondary title, a directory of topics of interest associated with the query information is constructed, and the directory of topics of interest is displayed in a channel page corresponding to the search extension tag.

[0257] In one or more embodiments, the page display module 101 is further configured to perform the following steps:

[0258] In response to a triggering operation on a search extension tag in the first result page, displaying a channel page corresponding to the search extension tag;

[0259] In the channel page corresponding to the search extension tag, a directory of topics of interest and a result return control associated with the query information are displayed;

[0260] The focus topic directory includes multiple focus topics associated with the query information and recommended terms corresponding to the multiple focus topics; the result return control is used to return from the channel page corresponding to the search extension page to the first result page.

[0261] In one or more embodiments, the page display module 101 is further configured to perform the following steps:

[0262] In response to a sliding operation on the first result page, updating displayed content on the first result page according to a sliding speed indicated by the sliding operation;

[0263] When the displayed content in the first result page includes the focus display area, hiding the displayed search expansion floating layer in the first result page;

[0264] The focus display area is used to display multiple focus topics associated with the query information and recommended terms corresponding to the multiple focus topics.

[0265] In one or more embodiments, the focus point display area also includes a detail viewing control;

[0266] The page display module 101 is also used to perform the following steps:

[0267] In response to a trigger operation on a detail viewing control in the focus display area, a channel page corresponding to the search extension tag in the first result page is displayed, and a directory of focus topics associated with the query information is displayed in the channel page corresponding to the search extension tag;

[0268] The focus topic catalog includes multiple focus topics contained in the focus point display area and recommended entries corresponding to the multiple focus topics.

[0269] In one or more embodiments, the page display module 101 is further configured to perform the following steps:

[0270] If it is detected that among the recommended terms of the plurality of focus topics associated with the query information, there are recommended terms that have been viewed, then in the focus point display area, the recommended terms that have been viewed and the recommended terms that have not been viewed are displayed separately.

[0271] In one or more embodiments, the data processing device further includes: a retrieval module 103, a term screening module 104, and a catalog generation module 105;

[0272] The search module 103 is used to obtain query information and search the query information in the index database; the index database includes multiple index information, one index information corresponds to multiple focus topics, and one focus topic includes multiple recommended candidate entries;

[0273] The retrieval module 103 is further configured to obtain, in the index database, a plurality of topics of interest corresponding to the target index information and a plurality of recommended candidate terms for each topic of interest corresponding to the target index information if target index information matching the query information is retrieved in the index database;

[0274] The term screening module 104 is used to screen multiple recommended candidate terms for each topic of interest corresponding to the target index information to obtain a recommended term for each topic of interest corresponding to the target search information;

[0275] The catalog generation module 105 is used to generate a catalog of topics of interest associated with the query information according to the multiple topics of interest corresponding to the target index information and the recommended terms of each topic of interest corresponding to the target search information.

[0276] In one or more embodiments, the index database includes an inverted index library and a semantic index library;

[0277] The search module 103 obtains the query information and searches the query information in the index database to perform the following steps:

[0278] Obtain query information, perform semantic analysis on the query information, and obtain a text semantic vector corresponding to the query information;

[0279] If a query statement matching the query information is retrieved in the inverted index library, the query statement matching the query information in the inverted index library is determined as target index information of the query information in the inverted index library;

[0280] If a semantic index matching the text semantic vector is retrieved in the semantic index library, the semantic index matching the text semantic vector in the semantic index library is determined as the target index information of the query information in the semantic index library.

[0281] In one or more embodiments, the entry screening module 104 is specifically configured to perform the following steps:

[0282] Obtaining a text semantic vector corresponding to the query information, and adding a plurality of recommended candidate words for each topic of interest corresponding to the target index information to the first candidate word set;

[0283] Obtaining the term semantic vector corresponding to each recommended candidate term in the first candidate term set, and obtaining the first similarity between the text semantic vector and the term semantic vector;

[0284] In the first candidate term set, the recommended candidate terms whose first similarity is less than the first similarity threshold are deleted to obtain a second candidate term set;

[0285] Obtaining a feedback index estimate value corresponding to each recommended candidate term in the second candidate term set, and obtaining a second similarity between each recommended candidate term in the second candidate term set according to the term semantic vector;

[0286] If in the second candidate term set, there exists a second similarity between the recommended candidate term a and the recommended candidate term b that is greater than the second similarity threshold, and the feedback index estimated value of the recommended candidate term a is greater than the feedback index estimated value of the recommended candidate term b, then the recommended candidate term b is deleted from the second candidate term set to obtain a third candidate term set;

[0287] Among the recommended candidate terms included in the third candidate term set, a recommended term for each topic of interest corresponding to the target search information is determined.

[0288] In one or more embodiments, the directory generation module 105 is specifically configured to perform the following steps:

[0289] Determine the estimated feedback index value of each topic of interest corresponding to the target index information according to the estimated feedback index value corresponding to the recommended term of each topic of interest;

[0290] According to the estimated value of the feedback indicator of each topic of interest, multiple topics of interest corresponding to the target index information are sorted to obtain multiple sorted topics of interest;

[0291] According to the estimated feedback indicator value corresponding to each recommended term of the concerned topic, the recommended terms of each concerned topic are sorted to obtain the sorted recommended terms corresponding to each concerned topic;

[0292] A directory of topics of interest associated with the query information is generated according to the sorted multiple topics of interest and the sorted recommended terms corresponding to each topic of interest.

[0293] In one or more embodiments, the data processing device further includes: a query statement collection module 106, a focus point generation module 107, a term clustering module 108, a topic deduplication module 109, a term deduplication module 110, and an adding module 111;

[0294] The query statement collection module 106 is used to obtain historical query statements and M historical candidate terms corresponding to the historical query statements; M is an integer greater than 1;

[0295] A focus generating module 107, configured to input the historical query statement into a language model, generate a first branch topic corresponding to the historical query statement through the language model, and generate a first branch term corresponding to the first branch topic through the language model;

[0296] The term clustering module 108 is used to perform clustering processing on the M historical candidate terms corresponding to the historical query sentence to obtain N candidate term groups, and generate a second branch topic for each candidate term group; N is a positive integer;

[0297] Topic deduplication module 109, used to perform deduplication processing on the first branch terms and the second branch topics to obtain the concerned topics corresponding to the historical query statements;

[0298] The term deduplication module 110 is used to perform deduplication processing on the M historical candidate terms and the first branch terms corresponding to the first branch topic to obtain the recommended candidate terms of the topic of interest corresponding to the historical query statement;

[0299] The adding module 111 is used to add the historical query statements, the concerned topics corresponding to the historical query statements, and the recommended candidate terms of the concerned topics corresponding to the historical query statements to the index database.

[0300] In one or more embodiments, the focus generation module 107 inputs the historical query statement into the language model, and generates a first branch topic corresponding to the historical query statement through the language model, so as to perform the following steps:

[0301] Input the historical query sentences into the language model, and convert the historical query sentences into historical embedding representations through the embedding layer in the language model;

[0302] Through the encoding component in the language model, the historical embedding representation is encoded to obtain the historical encoding features corresponding to the historical query sentences;

[0303] Through the fully connected layer in the language model, the historical encoding features are transformed to generate the first branch topic corresponding to the historical query statement.

[0304] In one or more embodiments, the term clustering module 108 performs clustering processing on the M historical candidate terms corresponding to the historical query statement to obtain N candidate term groups, and generates a second branch topic for each candidate term group to perform the following steps:

[0305] Convert the M historical candidate terms corresponding to the historical query statements into historical text vectors, add the historical text vectors corresponding to the M historical candidate terms to the text vector set, and select the center vector t in the text vector set. i ; i is a positive integer less than or equal to N;

[0306] Get each historical text vector and the center vector t in the text vector set i The third similarity between them, if the historical text vector G in the text vector set j With the center vector t i The third similarity between them is the largest, so the historical text vector G j Add to the center vector t i The pending cluster C i ; j is a positive integer less than or equal to M;

[0307] According to the cluster C to be processed i The historical text vector contained in the update center vector t i , until the cluster to be processed C i The updated center vector t i and the center vector t before update i If the same, according to the cluster C to be processed i Obtain N candidate entry groups;

[0308] The historical candidate terms in each candidate term group are input into the language model, and the second branch topic corresponding to each candidate term group is generated through the language model.

[0309] According to an embodiment of the present application, the above Figure 2 The relevant steps involved in the data processing method shown can be represented by Fig.12 The data processing device 1 shown in the figure is executed by each module. For example, Figure 2 Step S101 and step S102 shown in FIG. Fig.12 The page display module 101 shown is used to execute, Figure 2 Step S103 shown can be performed by Fig.12 The directory display module 102 shown is used to perform the following operations.

[0310] According to one embodiment of the present application, Fig.12 The various modules in the data processing device 1 shown can be separately or all combined into one or several modules to constitute, or one (some) of the modules can be further divided into at least two smaller units in function, and the same operation can be achieved without affecting the realization of the technical effects of the embodiments of the present application. The above modules are divided based on logical functions. In practical applications, the function of one module can also be implemented by at least two units, or the functions of at least two modules can be implemented by one module. In other embodiments of the present application, the data processing device 1 may also include other modules or units. In practical applications, these functions can also be implemented with the assistance of other modules, and can be implemented by the collaboration of at least two modules.

[0311] In an embodiment of the present application, when an object enters query information in a search box, a search operation for the query information can be responded to, and a first result page corresponding to the query information can be displayed, and the search results and a search expansion floating layer corresponding to the query information can be displayed in the first result page. The search expansion floating layer is independently displayed on the first result page, and the tab bar of the first result page may include a search expansion tag; both the search expansion floating layer and the search expansion tag can jump to the channel page corresponding to the search expansion tag, and the directory of topics of interest associated with the query information is displayed in the channel page corresponding to the search expansion tag. Since the directory of topics of interest contains multiple topics of interest associated with the query information, as well as the recommended terms under each topic of interest, and the directory of topics of interest is displayed in the channel page corresponding to the search expansion tag; by establishing a clear "outline directory-style" perception of all recommended terms associated with the query information, that is, the directory of topics of interest, it is beneficial for the object to click and view the recommended terms of interest in the topic directory in turn, which can improve the click-through rate of the directory of topics of interest, and thus increase the search traffic. By generating the concerned topics corresponding to the historical query statements and the recommended terms contained in each concerned topic through the generation method and the clustering method, the recommended terms associated with each historical query statement can be clustered in an organized manner, which can improve the coverage of the recommended terms.

[0312] See also Fig.13 , Fig.13 Schematic diagram of the structure of a computer device provided in an embodiment of the present application. Fig.13 As shown, the computer device 1000 may be a terminal device, for example, Figure 1 The terminal device 10a in the corresponding embodiment may also be a server, for example, Figure 1The server 10d in the corresponding embodiment will not be limited here. For ease of understanding, this application takes a computer device as an example of a terminal device. The computer device 1000 may include: a processor 1001, a network interface 1004 and a memory 1005. In addition, the computer device 1000 may also include: a user interface 1003, and at least one communication bus 1002. Among them, the communication bus 1002 is used to realize the connection and communication between these components. Among them, the user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory, or it may be a non-volatile memory (non-volatile memory), such as at least one disk storage. The memory 1005 may optionally also be at least one storage device located away from the aforementioned processor 1001. As Fig.13 As shown, the memory 1005 as a computer-readable storage medium may include an operating system, a network communication module, a user interface module, and a device control application program.

[0313] The network interface 1004 in the computer device 1000 can also provide a network communication function, and the optional user interface 1003 can also include a display screen (Display) and a keyboard (Keyboard). Fig.13 In the computer device 1000 shown, the network interface 1004 can provide a network communication function; the user interface 1003 is mainly used to provide an input interface for the user; and the processor 1001 can be used to call the device control application stored in the memory 1005 to achieve:

[0314] In response to a search operation for the query information, display a first result page corresponding to the query information;

[0315] On the first result page, the search results corresponding to the query information and the search expansion floating layer associated with the query information are displayed; the search expansion floating layer is displayed independently on the first result page, and the tab bar of the first result page includes a search expansion tag;

[0316] In response to a detail viewing operation on the search extension floating layer, a directory of topics of interest associated with the query information is displayed in a channel page corresponding to the search extension tag.

[0317] It should be understood that the computer device 1000 described in the embodiment of the present application can execute the above Figure 2 , Fig.10 The description of the data processing method in any of the embodiments may also be performed as described above. Fig.12The description of the data processing device 1 in the corresponding embodiment will not be repeated here. In addition, the description of the beneficial effects of the same method will not be repeated here either.

[0318] In addition, it should be pointed out here that: the embodiment of the present application also provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program executed by the data processing device 1 mentioned above, and the computer program includes computer instructions. When the processor executes the computer instructions, it can execute the above-mentioned Figure 2 , Fig.10 The description of the data processing method in any of the embodiments will not be repeated here. In addition, the description of the beneficial effects of the same method will not be repeated. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc. For technical details not disclosed in the computer-readable storage medium embodiment involved in this application, please refer to the description of the method embodiment of this application. As an example, program instructions can be deployed on a computer device for execution, or on multiple computer devices located at one location, or on multiple computer devices distributed at multiple locations and interconnected by a communication network. Multiple computer devices distributed at multiple locations and interconnected by a communication network can constitute a blockchain system.

[0319] In addition, it should be noted that: the embodiment of the present application also provides a computer program product, which may include a computer program, and the computer program may be stored in a computer-readable storage medium. The processor of the computer device reads the computer program from the computer-readable storage medium, and the processor may execute the computer program, so that the computer device performs the above Figure 2 , Fig.10 The description of the data processing method in any of the embodiments will not be repeated here. In addition, the description of the beneficial effects of the same method will not be repeated. For technical details not disclosed in the computer program product or computer program embodiment involved in the present application, please refer to the description of the method embodiment of the present application.

[0320] The terms "first", "second", etc. in the description, claims, and drawings of the embodiments of the present application are used to distinguish different media contents, rather than to describe a specific order. In addition, the term "comprising" and any variation thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product, or equipment comprising a series of steps or units is not limited to the listed steps or modules, but may optionally include steps or modules that are not listed, or may optionally include other step units inherent to these processes, methods, devices, products, or equipment.

[0321] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0322] The method and related apparatus provided by the embodiment of the present application are described with reference to the method flow chart and / or structural diagram provided by the embodiment of the present application. Specifically, each process and / or box in the method flow chart and / or structural diagram, as well as the combination of the processes and / or boxes in the flow chart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the process in the process. Figure 1 Schematic diagram of one or more processes and / or structures Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the functions specified in the process. Figure 1 Schematic diagram of one or more processes and / or structures Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process in the process. Figure 1 A flow or multiple flows and / or structures illustrate the steps of the functions specified in one block or multiple blocks.

[0323] In the embodiments of the present application, the term "module" or "unit" refers to a computer program or a part of a computer program with a predetermined function, and works together with other related parts to achieve a predetermined goal, and can be implemented in whole or in part by using software, hardware (such as processing circuits or memories) or a combination thereof. Similarly, a processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be part of an overall module or unit that includes the function of the module or unit.

[0324] The above disclosure is only the preferred embodiment of the present application, which certainly cannot be used to limit the scope of rights of the present application. Therefore, equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. A data processing method, characterized in that: include: In response to a search operation for query information, display a first result page corresponding to the query information; In the first result page, the search results corresponding to the query information and the search expansion floating layer associated with the query information are displayed; The search expansion floating layer is independently displayed on the first result page, and the tab bar of the first result page includes a search expansion tag; In response to a detail viewing operation on the search expansion floating layer, a directory of topics of interest associated with the query information is displayed in a channel page corresponding to the search expansion tag.

2. The method according to claim 1, characterized in that The method further comprises: In the entry area of ​​the search expansion floating layer, the recommended entries of the topic of interest associated with the query information are played in rotation; In response to a trigger operation on the term area, determining the recommended term displayed in the term area as a target term, and displaying a second result page corresponding to the target term; In the second result page, the search results and directory viewing controls corresponding to the target entry are displayed; the search extension tab in the tab bar of the second result page is in a hidden state.

3. The method according to claim 2, characterized in that The method further comprises: In response to a triggering operation on the directory viewing control in the second result page, displaying a directory of topics of interest associated with the query information in a channel page corresponding to the search extension tag; The target entry in the concerned topic catalog is highlighted, and in the concerned topic catalog, an extended entry is added to the concerned topic to which the target entry belongs.

4. The method according to claim 3, characterized in that In the concerned topic catalog, adding an extended entry for the concerned topic to which the target entry belongs includes: According to the display position of the target term in the channel page corresponding to the search extension tag, displaying the term connection animation in the focus topic directory; the term connection animation is used to rotate the extended terms of the focus topic to which the target term belongs; When the entry connection animation is finished, the extended entries included in the entry connection animation are expanded and displayed in the focus topic directory, and the display positions of multiple focus topics in the focus topic directory and the recommended entries of the multiple focus topics are adjusted.

5. The method according to claim 1, characterized in that In response to the detail viewing operation of the search expansion floating layer, displaying a directory of topics of interest associated with the query information in a channel page corresponding to the search expansion tag includes: In response to a detail viewing operation on the search expansion floating layer, a plurality of focus topics associated with the query information and a recommended entry corresponding to each focus topic are obtained; Determine multiple topics of interest associated with the query information as first-level titles, and determine recommended terms corresponding to each topic of interest as second-level titles; A directory of topics of interest associated with the query information is constructed according to the primary title and the secondary title, and the directory of topics of interest is displayed in a channel page corresponding to the search extension tag.

6. The method according to claim 1, characterized in that The method further comprises: In response to a triggering operation on a search extension tag in the first result page, displaying a channel page corresponding to the search extension tag; In the channel page corresponding to the search extension tag, displaying a directory of topics of interest and a result return control associated with the query information; The focus topic directory includes multiple focus topics associated with the query information and recommended terms corresponding to the multiple focus topics; the result return control is used to return from the channel page corresponding to the search extension page to the first result page.

7. The method according to claim 1, characterized in that The method further comprises: In response to a sliding operation on the first result page, updating displayed content on the first result page according to a sliding speed indicated by the sliding operation; When the displayed content in the first result page includes a focus display area, hiding and displaying the search expansion floating layer in the first result page; The focus display area is used to display multiple focus topics associated with the query information and recommended terms corresponding to the multiple focus topics.

8. The method according to claim 7, characterized in that The focus point display area also includes a detail viewing control; The method further comprises: In response to a triggering operation on a detail viewing control in the focus display area, displaying a channel page corresponding to the search extension tag in the first result page, and displaying a directory of focus topics associated with the query information in the channel page corresponding to the search extension tag; The focus topic catalog includes a plurality of focus topics contained in the focus point display area and recommended entries corresponding to the plurality of focus topics.

9. The method according to claim 7, characterized in that: The method further comprises: If it is detected that among the recommended terms of the plurality of focus topics associated with the query information, there are recommended terms that have been viewed, then in the focus point display area, the recommended terms that have been viewed and the recommended terms that have not been viewed are displayed separately.

10. The method according to any one of claims 1 to 9, characterized in that: The method further comprises: Acquire the query information, and search the query information in an index database; the index database includes a plurality of index information, one index information corresponds to a plurality of focus topics, and one focus topic includes a plurality of recommended candidate entries; If target index information matching the query information is retrieved in the index database, a plurality of topics of interest corresponding to the target index information and a plurality of recommended candidate entries for each topic of interest corresponding to the target index information are obtained in the index database; Screening multiple recommended candidate terms for each topic of interest corresponding to the target index information to obtain a recommended term for each topic of interest corresponding to the target search information; A directory of topics of interest associated with the query information is generated according to the multiple topics of interest corresponding to the target index information and the recommended terms of each topic of interest corresponding to the target search information.

11. The method according to claim 10, characterized in that The index database includes an inverted index library and a semantic index library; The obtaining of the query information and searching the query information in an index database includes: Acquire the query information, perform semantic analysis on the query information, and obtain a text semantic vector corresponding to the query information; If a query statement matching the query information is retrieved in the inverted index library, the query statement matching the query information in the inverted index library is determined as target index information of the query information in the inverted index library; If a semantic index matching the text semantic vector is retrieved in the semantic index library, the semantic index matching the text semantic vector in the semantic index library is determined as target index information of the query information in the semantic index library.

12. The method according to claim 10, characterized in that The step of screening a plurality of recommended candidate terms for each topic of interest corresponding to the target index information to obtain a recommended term for each topic of interest corresponding to the target search information includes: Acquire a text semantic vector corresponding to the query information, and add a plurality of recommended candidate words for each topic of interest corresponding to the target index information to a first candidate word set; Obtaining a term semantic vector corresponding to each recommended candidate term in the first candidate term set, and obtaining a first similarity between the text semantic vector and the term semantic vector; In the first candidate entry set, the recommended candidate entries whose first similarity is less than the first similarity threshold are deleted to obtain a second candidate entry set; Obtaining a feedback index estimate value corresponding to each recommended candidate term in the second candidate term set, and obtaining a second similarity between each recommended candidate term in the second candidate term set according to the term semantic vector; If in the second candidate term set, there exists a second similarity between the recommended candidate term a and the recommended candidate term b that is greater than the second similarity threshold, and the feedback index estimated value of the recommended candidate term a is greater than the feedback index estimated value of the recommended candidate term b, then the recommended candidate term b is deleted from the second candidate term set to obtain a third candidate term set; Among the recommended candidate terms included in the third candidate term set, a recommended term for each topic of interest corresponding to the target search information is determined.

13. The method according to claim 10, characterized in that The step of generating a catalog of topics of interest associated with the query information according to the multiple topics of interest corresponding to the target index information and the recommended terms of each topic of interest corresponding to the target search information includes: Determining the estimated feedback index value of each topic of interest corresponding to the target index information according to the estimated feedback index value corresponding to the recommended term of each topic of interest; Sorting the multiple topics of interest corresponding to the target index information according to the estimated feedback index value of each topic of interest, to obtain the sorted multiple topics of interest; According to the estimated feedback index value corresponding to the recommended terms of each topic of interest, the recommended terms of each topic of interest are sorted to obtain sorted recommended terms corresponding to each topic of interest; A directory of topics of interest associated with the query information is generated according to the sorted multiple topics of interest and the sorted recommended terms corresponding to each topic of interest.

14. The method according to claim 10, characterized in that The method further comprises: Obtain a historical query statement and M historical candidate entries corresponding to the historical query statement; M is an integer greater than 1; Inputting the historical query statement into a language model, generating a first branch topic corresponding to the historical query statement through the language model, and generating a first branch term corresponding to the first branch topic through the language model; Perform clustering processing on the M historical candidate terms corresponding to the historical query sentence to obtain N candidate term groups, and generate a second branch topic for each candidate term group; N is a positive integer; Deduplication processing is performed on the first branch terms and the second branch topics to obtain the concerned topics corresponding to the historical query statements; Deduplication processing is performed on the M historical candidate terms and the first branch terms corresponding to the first branch topic to obtain recommended candidate terms for the topic of interest corresponding to the historical query statement; The historical query statement, the concerned topic corresponding to the historical query statement, and the recommended candidate terms of the concerned topic corresponding to the historical query statement are added to the index database.

15. The method according to claim 14, characterized in that The step of inputting the historical query statement into a language model and generating a first branch topic corresponding to the historical query statement through the language model includes: Inputting the historical query sentence into a language model, and converting the historical query sentence into a historical embedding representation through an embedding layer in the language model; The historical embedding representation is encoded by the encoding component in the language model to obtain the historical encoding features corresponding to the historical query statement; The historical encoding features are subjected to feature transformation processing through the fully connected layer in the language model to generate a first branch topic corresponding to the historical query statement.

16. The method according to claim 14, characterized in that The clustering process is performed on the M historical candidate terms corresponding to the historical query sentence to obtain N candidate term groups, and a second branch topic is generated for each candidate term group, including: The M historical candidate terms corresponding to the historical query sentence are converted into historical text vectors, the historical text vectors corresponding to the M historical candidate terms are added to the text vector set, and the center vector t is selected from the text vector set. i ; i is a positive integer less than or equal to N; Get each historical text vector in the text vector set and the center vector t i The third similarity between them is that if the historical text vector G in the text vector set j With the center vector t i The third similarity between them is the largest, then the historical text vector G j Add to the center vector t i The pending cluster C i ; j is a positive integer less than or equal to M; According to the cluster C to be processed i The historical text vector contained in , updates the central vector t i , until the to-be-processed cluster C i The updated center vector t i and the center vector t before update i If the same, according to the to-be-processed cluster C i Obtain N candidate entry groups; The historical candidate terms in each candidate term group are input into a language model, and the second branch topic corresponding to each candidate term group is generated by the language model.

17. A data processing device, characterized in that: include: A page display module, configured to respond to a search operation for query information and display a first result page corresponding to the query information; The page display module is further used to display the search results corresponding to the query information and the search expansion floating layer associated with the query information in the first result page; The search expansion floating layer is independently displayed on the first result page, and the tab bar of the first result page includes a search expansion tag; The directory display module is used to respond to the detail viewing operation on the search expansion floating layer and display the concerned topic directory associated with the query information in the channel page corresponding to the search expansion tag.

18. A computer device, characterized in that: including memory and processor; The memory is connected to the processor, the memory is used to store a computer program, and the processor is used to call the computer program so that the computer device executes the method according to any one of claims 1 to 16.

19. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded and executed by a processor, so that a computer device having the processor executes the method according to any one of claims 1 to 16.

20. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 16 when being executed by a processor.

Images