Content search method, device, electronic device and storage medium

Abstract

The present disclosure relates to a content search method, device, electronic device and storage medium, and relates to the field of data processing technology. The method includes: obtaining a search term and determining multiple search results corresponding to the search term; reading the correspondence between the pre-stored search term and at least two types of search results; the correspondence is determined based on the historical search results whose user satisfaction value and display times corresponding to the search term meet preset conditions; according to the correspondence, obtaining at least two types of target search results from multiple search results; and displaying the target search results.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to a content search method, device, electronic device and storage medium. Background Art

[0002] In the related art, most search systems will introduce a posteriori models such as clicks. Such search systems are more inclined to return content that meets the search needs of mainstream users in both the recall and sorting stages. For example, in a certain application scenario, the search term "Legend" is a search term where most users want to search for game legends. In this scenario, when searching for "Legend", the search system is more inclined to return content about game legends in both the recall and sorting stages.

[0003] Obviously, this behavior of focusing only on the search needs of mainstream users will result in the search needs of other users not being met, which in turn leads to a very poor search experience for other users. For example, for users who want to search for song legends, the search experience will be very poor. Summary of the invention

[0004] The present disclosure provides a content search method, device, electronic device and storage medium in a search scenario, so as to at least solve the problem that the search system in the related art only focuses on the search needs of mainstream users, resulting in the search needs of other users not being met, and thus resulting in a very poor search experience for other users. The technical solution of the present disclosure is as follows:

[0005] According to a first aspect of an embodiment of the present disclosure, a content search method is provided, comprising: obtaining a search term, and determining a plurality of search results corresponding to the search term; reading a pre-stored correspondence between the search term and at least two types of search results; the correspondence is determined based on historical search results whose user satisfaction value and display times corresponding to the search term meet preset conditions; based on the correspondence, obtaining at least two types of target search results from the plurality of search results; and displaying the target search results.

[0006] In a possible implementation, the content search method further includes: obtaining historical search results corresponding to the search term within a preset time period, and user behavior data of the historical search results; aggregating the user behavior data of each historical search result of the search term to obtain a user satisfaction value of each historical search result; obtaining historical search results whose user satisfaction value is greater than a first threshold and whose number of displays is greater than a second threshold to obtain samples to be marked; determining the type of each historical search result in the sample to be marked to obtain a correspondence between the search term and at least two types of search results.

[0007] In another possible implementation, the determining the type of each search result in the sample to be labeled includes: constructing a graph structure based on the sample to be labeled and a graph embedding model; the graph structure includes a first node for representing the historical search results in the sample to be labeled, a second node for representing the search term, and a connection relationship between the first node and the second node; determining the type of each historical search result in the sample to be labeled based on the graph structure. In another possible implementation, the determining the type of each historical search result in the sample to be labeled based on the graph structure includes: determining a maximal connected subgraph of the graph structure; obtaining a maximal connected subgraph with a node number less than a preset threshold to obtain a subgraph to be labeled; the node includes the first node and the second node; determining the type of the first node in the subgraph to be labeled.

[0008] In another possible implementation, determining the type of each historical search result in the sample to be marked based on the graph structure also includes: determining the embedding vector of the first node in the graph structure; determining the first node to be marked based on the embedding vector and the first node in the subgraph to be marked; and determining the type of the first node to be marked based on a clustering algorithm.

[0009] In another possible implementation, the user behavior data of each search result of the historical search term is aggregated to obtain a user satisfaction value for each historical search result, including: determining the user satisfaction of the historical search results based on the aggregated user behavior data; and determining the user satisfaction value of the historical search results based on the user satisfaction.

[0010] In another possible implementation, determining the user satisfaction of the historical search results based on the aggregated user behavior data includes: determining the user satisfaction of the historical search results based on each type of user behavior data of each user for the historical search results, and a weight coefficient of each type of user behavior data.

[0011] In another possible implementation, the user satisfaction value of the historical search results is determined based on the user satisfaction, including: determining the user satisfaction rate of the historical search results based on the user satisfaction and the number of times the historical search results are played; determining the user satisfaction confidence of the historical search results based on the number of times the historical search results are played; and determining the user satisfaction value of the historical search results based on the user satisfaction rate and the user satisfaction confidence.

[0012] According to a second aspect of an embodiment of the present disclosure, a content search device is provided, comprising: an acquisition module, configured to execute acquisition of a search term and determine a plurality of search results corresponding to the search term; a reading module, configured to execute reading of a pre-stored correspondence between the search term and at least two types of search results; the correspondence is determined based on historical search results whose user satisfaction value and display times corresponding to the search term meet preset conditions; the acquisition module is further configured to execute acquisition of at least two types of target search results from the plurality of search results based on the correspondence; and a display module, configured to execute display of the target search results.

[0013] In a possible implementation, the device also includes a configuration module, which is configured to perform the following operations: obtaining historical search results corresponding to the search term within a preset time period, and user behavior data of the historical search results; aggregating the user behavior data of each historical search result of the search term to obtain a user satisfaction value of each historical search result; obtaining historical search results whose user satisfaction value is greater than a first threshold and whose number of displays is greater than a second threshold to obtain samples to be marked; determining the type of each historical search result in the sample to be marked to obtain a correspondence between the search term and at least two types of search results.

[0014] In another possible implementation, the configuration module is specifically configured to perform: constructing a graph structure based on the sample to be labeled and the graph embedding model; the graph structure includes a first node for representing the historical search results in the sample to be labeled, a second node for representing the search term, and a connection relationship between the first node and the second node; and determining the type of each historical search result in the sample to be labeled based on the graph structure.

[0015] In another possible implementation, the configuration module is specifically configured to execute: determining a maximal connected subgraph of the graph structure; obtaining a maximal connected subgraph whose number of nodes is less than a preset threshold to obtain a subgraph to be marked; the nodes include the first node and the second node; and determining the type of the first node in the subgraph to be marked.

[0016] In another possible implementation, the configuration module is specifically configured to perform: determining an embedding vector of the first node in the graph structure; determining the first node to be marked based on the embedding vector and the first node in the subgraph to be marked; and determining a type of the first node to be marked based on a clustering algorithm.

[0017] In another possible implementation, the configuration module is specifically configured to execute: determining user satisfaction of the historical search results based on the aggregated user behavior data; and determining a user satisfaction value of the historical search results based on the user satisfaction.

[0018] In another possible implementation, the configuration module is specifically configured to execute: determining user satisfaction with the historical search results based on each user behavior data of each user on the historical search results and a weight coefficient of each user behavior data.

[0019] In another possible implementation, the configuration module is specifically configured to perform: determining a user satisfaction rate of the historical search results based on the user satisfaction and the number of times the historical search results are played; determining a user satisfaction confidence level of the historical search results based on the number of times the historical search results are played; and determining a user satisfaction value of the historical search results based on the user satisfaction rate and the user satisfaction confidence level.

[0020] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute instructions to implement the content search method of the above-mentioned first aspect and any possible implementation manner thereof.

[0021] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the content search method in the above-mentioned first aspect and any possible implementation manner thereof.

[0022] According to a fifth aspect of an embodiment of the present disclosure, a computer program product is provided. The computer program product includes computer instructions. When the computer instructions are executed on an electronic device, the electronic device executes the content search method of the first aspect and any possible implementation manner thereof.

[0023] The technical solution provided by the embodiments of the present disclosure brings at least the following beneficial effects: after determining the multiple search results corresponding to the search term, the pre-stored correspondence between the search term and at least two types of search results is read, and according to the correspondence, at least two types of target search results are obtained from the multiple search results, and the target search results are displayed. Since the correspondence is determined based on the historical search results that meet the preset conditions for the user satisfaction value and the number of times displayed corresponding to the search term, that is, the at least two types of search results corresponding to the search term are the types corresponding to the search results required by the user account, therefore, by displaying at least two types of target search results, the search needs of more user accounts are met, avoiding the situation in which the search system in the prior art only focuses on the search needs of mainstream users and ignores the search needs of other users, resulting in the search needs of other users not being met, thereby improving the search experience of most users.

[0024] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description are used to explain the principles of the present disclosure, and do not constitute improper limitations on the present disclosure.

[0026] Figure 1 is a flow chart showing a content search method according to an exemplary embodiment;

[0027] Figure 2 is a flow chart showing another content search method according to an exemplary embodiment;

[0028] Figure 3 is a flow chart showing another content search method according to an exemplary embodiment;

[0029] Figure 4 is a flow chart showing another content search method according to an exemplary embodiment;

[0030] Figure 5 is a flow chart showing another content search method according to an exemplary embodiment;

[0031] Figure 6 is a block diagram of a content search device according to an exemplary embodiment;

[0032] Figure 7 It is a block diagram of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION

[0033] In order to enable ordinary persons in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings.

[0034] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0035] Before introducing the content search method provided by the present disclosure in detail, the application scenarios and implementation environments involved in the present disclosure are briefly introduced.

[0036] First, the application scenarios involved in the present disclosure are briefly introduced.

[0037] Ideally, a good search system will return as many search results as possible based on the search terms entered by the user. However, returning only relevant search results may not always fully meet the needs of users. This is because, in search scenarios, the search terms entered by users usually have multiple meanings. Accordingly, search results corresponding to search terms with multiple meanings also have multiple types. If only relevant search results are returned without considering the type of search results required by users, it will be difficult to truly meet the search needs of users.

[0038] For example, the search term "Legend" has at least two meanings: game legend and song legend. Correspondingly, the search results corresponding to the search term "Legend" also have at least two types, one of which is a game type and the other is a song type.

[0039] Most search systems in the prior art will introduce a posteriori models such as clicks. Such search systems are more inclined to return content that meets the search needs of mainstream users in both the recall and sorting stages. For example, in a certain application scenario, the search term "Legend" is a search term where most users want to search for game legends. In this scenario, when searching for "Legend", the search system is more inclined to return content about game legends in both the recall and sorting stages.

[0040] Obviously, this behavior of focusing only on the search needs of mainstream users will result in the search needs of other users not being met, which in turn leads to a very poor search experience for other users. For example, for users who want to search for song legends, the search experience will be very poor.

[0041] In view of the above problems, the present disclosure provides a content search method, which reads the pre-stored correspondence between the search term and at least two types of search results after determining the multiple search results corresponding to the search term, and obtains at least two types of target search results from the multiple search results according to the correspondence, and displays the target search results. Since the correspondence is determined based on the historical search results that meet the preset conditions for the user satisfaction value and the number of times displayed corresponding to the search term, that is, the at least two types of search results corresponding to the search term are the types corresponding to the search results required by the user account, therefore, by displaying at least two types of target search results, it is achieved to meet the search needs of more user accounts, avoiding the situation in which the search system in the prior art only focuses on the search needs of mainstream users and ignores the search needs of other users, resulting in the search needs of other users not being met, thereby improving the search experience of most users.

[0042] Next, the implementation environment (implementation architecture) involved in the present disclosure is briefly introduced below.

[0043] The content search method provided in the embodiments of the present disclosure can be applied to electronic devices. The electronic device can be a terminal device or a server. The terminal device can be a smart phone, a tablet computer, a handheld computer, a vehicle terminal, a desktop computer, a laptop computer, etc. The server can be any server or server cluster, which is not limited in the present disclosure.

[0044] In addition, it should be noted that the user information involved in this disclosure (including but not limited to user device information, user personal information, user account information, etc.) is all information authorized by the user or fully authorized by all parties.

[0045] For ease of understanding, the content search method provided by the present disclosure is specifically introduced below with reference to the accompanying drawings.

[0046] Figure 1 It is a flowchart showing a content search method according to an exemplary embodiment. The content search method is used in an electronic device and includes S101-S104.

[0047] In S101, a search term is obtained, and a plurality of search results corresponding to the search term are determined.

[0048] Optionally, after the electronic device obtains the search term, the search system recalls multiple search results based on the search term.

[0049] For example, the electronic device obtains the search term "legend", and the search system recalls based on "legend" and obtains 30 search results.

[0050] Optionally, the search term may be a search term entered by a user account.

[0051] For example, the user account wants to know the relevant information of the song legend, so the user account enters the search word "legend" in the search bar. At this time, the electronic device obtains the search word "legend".

[0052] In S102, the correspondence between pre-stored search terms and at least two types of search results is read; the correspondence is determined based on historical search results whose user satisfaction values ​​and display times corresponding to the search terms meet preset conditions.

[0053] Optionally, the correspondence between the search term and the at least two types of search results may be pre-stored in a local cache of the electronic device. Alternatively, the correspondence between the search term and the at least two types of search results may be stored in other electronic devices.

[0054] Optionally, after determining a plurality of search results corresponding to a search term, the electronic device reads a pre-stored correspondence between the search term and at least two types of search results to determine a search result to be displayed.

[0055] Optionally, the pre-stored corresponding relationship is determined based on the target historical search results corresponding to the search term, and the user satisfaction value and the number of times displayed of the target search results respectively meet the preset conditions.

[0056] In one embodiment, the electronic device obtains all search results corresponding to each search term within a preset time period, selects historical search results whose user satisfaction value and display times meet preset conditions, and uses them as target historical search results. The type corresponding to each target search result is determined, thereby determining at least two types of search results corresponding to the search term.

[0057] Exemplarily, the preset condition corresponding to the user satisfaction value may be a first threshold value. For example, the first threshold value may be 0.05, and when the user satisfaction value of the historical search result is greater than 0.05, it is determined that the user satisfaction value of the historical search result meets the preset condition.

[0058] Exemplarily, the preset condition corresponding to the number of times displayed may be a second threshold value. For example, the second threshold value may be 0, and when the number of times the historical search results are displayed is greater than 0, it is determined that the number of times the historical search results are displayed meets the preset condition.

[0059] It should be noted that the user satisfaction value of the search result is used to represent the user account's satisfaction with the search result. That is, the larger the satisfaction value is, the more satisfied the user account is with the search result.

[0060] By determining at least two types of search results corresponding to a search term using historical search results whose user satisfaction values ​​and display times respectively meet preset conditions, it is possible to filter out the types of search results that satisfy the user account in the historical search records of the search term.

[0061] For example, for the search term "Legend", there are 20 historical search results whose user satisfaction value and display times meet the preset conditions. There are two types of these 20 historical search results, one is song legends and the other is game legends. Therefore, the two types of search results corresponding to the search term "Legend" are song legends and game legends.

[0062] In S103, at least two types of target search results are obtained from the multiple search results according to the corresponding relationship.

[0063] Optionally, the electronic device obtains at least two similar target search results from the multiple search results according to a pre-stored corresponding relationship.

[0064] Exemplarily, as described above, the two types of search results corresponding to the search term "Legend" are song legends and game legends. The electronic device obtains two types of target search results, song legends and game legends, from the multiple search results.

[0065] In S104, the target search results are displayed.

[0066] Optionally, the electronic device displays the target search results on a corresponding content display interface of the search system.

[0067] Exemplarily, after the user account enters a search term in the search bar, the target display result can be displayed on the page below the search bar.

[0068] Exemplarily, after the user account enters the search word "legend" in the search bar, two types of target search results, song legends and game legends, are displayed below the search bar.

[0069] In the above embodiment, after determining the multiple search results corresponding to the search term, the pre-stored correspondence between the search term and at least two types of search results is read, and according to the correspondence, at least two types of target search results are obtained from the multiple search results, and the target search results are displayed. Since the correspondence is determined based on the historical search results that meet the preset conditions for the user satisfaction value and the number of times displayed corresponding to the search term, that is, the at least two types of search results corresponding to the search term are the types corresponding to the search results required by the user account, therefore, by displaying at least two types of target search results, the search needs of more user accounts are met, avoiding the situation in which the search system in the prior art only focuses on the search needs of mainstream users and ignores the search needs of other users, resulting in the search needs of other users not being met, thereby improving the search experience of most users.

[0070] In one possible implementation, combining Figure 1 ,like Figure 2 As shown, the content search method further includes S105-S108. S105-S108 are used in electronic devices.

[0071] In S105, historical search results corresponding to the search term within a preset time period and user behavior data of the historical search results are obtained.

[0072] Optionally, the preset time period is used to represent a historical time period before the current moment.

[0073] Exemplarily, the preset time period may be 30 days before the current moment.

[0074] It can be understood that the historical search results corresponding to the search terms and the user behavior data of the historical search results are all historical data.

[0075] Optionally, the user behavior includes at least one or more of clicking, playing, liking, and following the author.

[0076] Exemplarily, user behavior data is used to characterize the user account's behavior operations on historical search results, such as the user account clicking on historical search results, playing historical search results, liking historical search results, and following the author corresponding to the historical search results.

[0077] In one embodiment, the user behavior data includes at least one or more of the number of clicks, the number of plays, the number of likes, and the number of times an author is followed.

[0078] Exemplarily, the user behavior data is used to characterize the number of user behavior operations on historical search results, such as the number of clicks on historical search results, the number of plays on historical search results, and the number of follows on the author corresponding to the historical search results.

[0079] Optionally, the number of times the historical search results have been displayed may also be obtained.

[0080] Exemplarily, the number of times a historical search result is displayed is used to represent the number of times the historical search result is displayed on the display interface of the electronic device.

[0081] Optionally, the historical search results may be multimedia resources, such as text, pictures, videos, audio, mini-programs, or live broadcasts.

[0082] In S106, the user behavior data of each historical search result of the search term is aggregated to obtain a user satisfaction value of each historical search result.

[0083] Optionally, the electronic device aggregates the user behavior data of each historical search result of the search term, and obtains the user satisfaction value of each historical search result based on the aggregated user behavior data of each historical search result. In one embodiment, the user behavior data of each historical search result is aggregated by summing the user behavior data of the same type corresponding to each historical search result.

[0084] Exemplarily, the same type of user behavior data refers to the same user behavior of different user accounts on historical search results. For example, click operations performed by different user accounts on historical search results belong to the same type of user behavior data.

[0085] For example, the aggregation of user behavior data of each historical search result of a search term can be achieved by forming <search term, historical search result, aggregation (click), aggregation (play), aggregation (like), aggregation (follow user), other user behavior data>. Among them, other user behavior data can be short broadcast behavior, which can be characterized by a playback time of less than 3 seconds.

[0086] For example, user account A clicks on the first historical search result, user account B clicks on the first historical search result, and user account C clicks on the first historical search result. At this time, in the aggregated user behavior data of the first historical search result, the data corresponding to the click is 3 times.

[0087] In one embodiment, the electronic device determines the user satisfaction value of each historical search result based on the aggregated user behavior data of click (3 times), play (10 times), like (8 times), and follow the author (5 times) for the first historical search results.

[0088] Optionally, the aggregated user behavior data is used to characterize the user's attention to the search result. For example, the higher the aggregated click count, the higher the user's attention to the search result, and further the higher the user's demand for the search result.

[0089] By aggregating the user behavior data of the search results, it is possible to determine the attention received by the search results, and further determine the user's demand for the search results.

[0090] In S107, historical search results whose user satisfaction value is greater than a first threshold and whose display times are greater than a second threshold are obtained to obtain samples to be marked.

[0091] Optionally, after obtaining the user satisfaction value of each historical search result, the electronic device screens the user satisfaction values ​​to obtain historical search results with user satisfaction values ​​greater than a first threshold. After obtaining historical search results with user satisfaction values ​​greater than the first threshold, the electronic device screens historical search results with display times greater than a second threshold to obtain samples to be marked.

[0092] Optionally, the preset condition corresponding to the user satisfaction value is a first threshold value. If the user satisfaction value satisfies the preset condition, the user satisfaction value is greater than the first threshold value.

[0093] Exemplarily, the first threshold may be 0.05. When the user satisfaction value of the historical search results is greater than 0.05, it is determined that the user satisfaction value of the historical search results meets the preset condition.

[0094] Optionally, the preset condition corresponding to the number of times of display is a second threshold value. The number of times of display corresponding to the number of times of display satisfying the preset condition is greater than the second threshold value.

[0095] Exemplarily, the second threshold may be 0. When the number of times the historical search results are displayed is greater than 0, it is determined that the number of times the historical search results are displayed meets the preset condition.

[0096] Optionally, the user satisfaction value of the historical search results may be used to represent the user's satisfaction with the historical search results.

[0097] Optionally, each historical search result having a user satisfaction value greater than a first threshold and a display number greater than a second threshold is determined as a sample to be marked.

[0098] In one embodiment, each sample to be marked includes a search term, a historical search result corresponding to the search term, and aggregated user behavior data of the historical search result. Furthermore, the sample to be marked may also include a user satisfaction value and / or display times of the historical search result.

[0099] It can be understood that since a sample to be marked includes a historical search result, the user satisfaction value of the historical search result corresponds to the sample to be marked one by one, that is, one sample to be marked corresponds to one user satisfaction value. The historical search results are screened by the user satisfaction value and the number of times displayed to obtain the historical search results that meet the user account search requirements as the sample to be marked, so that when at least two types of search results corresponding to the search terms obtained according to the sample to be marked are obtained, the search result type that meets the user's search requirements can be obtained.

[0100] In S108, the type of each historical search result in the sample to be marked is determined, and a corresponding relationship between the search term and at least two types of search results is obtained.

[0101] Optionally, after the electronic device determines the type of each historical search result in the sample to be marked, it filters the search result types in the sample to be marked, determines at least two types of search results corresponding to the search term, and obtains a correspondence between the search term and the at least two types of search results.

[0102] For example, for the search term "Legend", there are 20 historical search results in the sample to be marked, and the electronic device determines the type of each historical search result. The search result type of the sample to be marked is screened, and it is determined that there are two types of search results in the sample to be marked: song legend and game legend, so as to obtain the corresponding relationship between the search term "Legend" and the two types of search results: song legend and game legend.

[0103] Optionally, the type of the historical search results may be determined by marking a type identifier for each historical search result in the sample to be labeled.

[0104] Exemplarily, historical search results of the same type have the same type identifier. For example, the historical search structure corresponding to the search term "legend" includes at least two types of search results: song legend and game legend. When determining the type of search results, the first identifier can be used to represent the historical search results corresponding to the song legend type, and the second identifier can be used to represent the historical search results corresponding to the game legend type.

[0105] In the above embodiment, the user behavior data of each historical search result of the search term is aggregated to obtain the user satisfaction value of each historical search result, and the historical search results with a user satisfaction value greater than a first threshold and a display number greater than a second threshold are used as samples to be marked, thereby obtaining the correspondence between the search term and at least two types of search results based on the type of each historical search result in the sample to be marked. Since the search result type corresponding to the search term is determined based on the historical search results that meet the conditions of user satisfaction value and display number, that is, it is determined based on the search results that meet the user's search needs, the search result type corresponding to the search term is the type that meets the user's search needs, thereby achieving the determination of multiple types of search results that meet the user's search needs.

[0106] In one possible implementation, combining Figure 2 ,like Figure 3 As shown, S108 includes S108a-S108b.

[0107] In S108a, a graph structure is constructed according to the samples to be labeled and the graph embedding model.

[0108] The graph structure includes a first node for representing historical search results in the sample to be marked, a second node for representing the search term, and a connection relationship between the first node and the second node.

[0109] Optionally, the electronic device constructs a graph structure based on a graph embedding model, with historical search results in the sample to be marked and search terms corresponding to the historical search results in the sample to be marked as nodes, and user satisfaction values ​​of the historical search results in the sample to be marked as the connection relationship between the historical search results and the search terms.

[0110] Optionally, the connection relationship between the first node and the second node is used to characterize a user satisfaction value of historical search results in the sample to be marked.

[0111] Optionally, the graph structure further includes a connection edge between the first node and the second node. The connection edge is used to represent the corresponding relationship between the historical search results and the search terms in the sample to be marked.

[0112] Exemplarily, there is a connection edge between the search term and the historical search results corresponding to the search term.

[0113] In one embodiment, the graph structure is composed of<V,E,rel> In structure, V represents the node set, E represents the edge set, and rel represents the connection relationship set.

[0114] Exemplarily, the node set V may be represented by a first node representing a search result and a second node representing a historical search term.

[0115] Exemplarily, the connection edge set E can be represented by the connection edges between the historical search results and the search terms. It can be understood that there are connection edges between the historical search results and the search terms belonging to the same sample to be labeled.

[0116] For example, the connection relationship set rel satisfies the following formula: int(100*satisfy score ), satisfying score Characterize the user satisfaction value of the historical search results in each sample to be labeled.

[0117] It should be noted that the INT() function rounds a real number (which can be a mathematical expression) down to the nearest integer.

[0118] In S108b, the type of each historical search result in the sample to be marked is determined according to the graph structure.

[0119] Optionally, after obtaining the graph structure, the electronic device obtains the type of each historical search result in the sample to be marked by determining the type of the node representing the historical search result in the graph structure.

[0120] Exemplarily, the type of the first node in the graph structure is determined according to a clustering algorithm, and the type of each historical search result in the sample to be marked is determined.

[0121] Optionally, the type of the first node in the graph structure is determined by a pre-trained machine learning model.

[0122] Exemplarily, the graph structure is input into a pre-trained machine learning model, and the pre-trained machine learning model labels the first node in the graph structure with a type identifier.

[0123] In the above embodiment, a graph structure is constructed by taking the historical search results in the sample to be marked and the search terms corresponding to the historical search results in the sample to be marked as nodes, and taking the user satisfaction values ​​of the historical search results in the sample to be marked as the connection relationship between the historical search results and the search terms. The historical search results in the sample to be marked, the search terms corresponding to the historical search results, and the user satisfaction values ​​of the historical search results are displayed through a graph network, so that the type of each historical search result in the sample to be marked is obtained by determining the type of the node in the graph structure. Compared with determining the type of the historical search result in each sample to be marked one by one, the amount of calculation is reduced, and the accuracy and efficiency of type determination are improved.

[0124] In one possible implementation, combining Figure 3 ,like Figure 4 As shown, S108b includes S108b1-S108b3.

[0125] In S108b1, a maximal connected subgraph of the graph structure is determined.

[0126] Optionally, the maximal connected graph is used to represent a subgraph without a connection point in the graph structure. That is, any two maximal connected subgraphs have no connection points in the graph structure. In other words, in the graph structure, there is no connecting edge between any two maximal connected subgraphs.

[0127] Exemplarily, the graph structure includes five subgraphs, namely, the first subgraph, the second subgraph, the third subgraph, the fourth subgraph, and the fifth subgraph. There is a connection point between the first subgraph and the second subgraph, that is, there is a connection edge between the first subgraph and the second subgraph, and there is no connection edge between the other subgraphs. In this case, the graph structure has the third subgraph, the fourth subgraph, and the fifth subgraph. Figure 3 A maximal connected subgraph.

[0128] In S108b2, a maximal connected subgraph with a node number less than a preset threshold is obtained to obtain a subgraph to be marked; the nodes include a first node and a second node.

[0129] Optionally, the preset threshold may be 5.

[0130] Exemplarily, as described above, the graph structure has a third subgraph, a fourth subgraph, and a second subgraph. Figure 3 maximal connected subgraphs. Among them, the number of nodes in the third subgraph (including the first node and the second node) is 20, the number of nodes in the fourth subgraph (including the first node and the second node) is 12, and the number of nodes in the fifth subgraph (including the first node and the second node) is 4. At this time, the electronic device determines the fifth subgraph as the subgraph to be marked.

[0131] It should be noted that the preset threshold value can be set according to specific scenarios, and can be greater than 5, or can also be less than 5. The present disclosure does not limit this.

[0132] By screening the maximally connected subgraphs with a preset threshold, the types of the first nodes in the subgraphs to be marked are nearly the same, and the type of the first node on the subgraphs to be marked can be directly determined, which simplifies the workload of determining the type of the first node.

[0133] In S108b3, the type of the first node in the subgraph to be marked is determined.

[0134] Optionally, the type of the first node may be determined by marking a type identifier for the first node, thereby achieving determination of the type of the search result.

[0135] For example, the type of the first node may be manually labeled, or may be automatically labeled using a pre-trained model.

[0136] In the above embodiment, the nodes in the graph structure are split by determining the maximal connected subgraph of the graph structure, and the subgraph to be marked with a single node type is filtered out from the maximal connected subgraph by a preset threshold, so that the first node in the subgraph to be marked can be directly marked by type, reducing the workload of determining the corresponding type of historical search results.

[0137] In one possible implementation, combining Figure 4 ,like Figure 5 As shown, S108b also includes S108b4-S108b6.

[0138] In S108b4, an embedding vector of a first node in the graph structure is determined.

[0139] Optionally, when determining the embedding vector of each first node in the graph structure, a graph embedding algorithm may be used for calculation.

[0140] Exemplarily, the graph embedding algorithm may use PyTorch BigGraph.

[0141] In S108b5, the first node to be marked is determined according to the embedding vector and the first node in the subgraph to be marked.

[0142] Optionally, the first node to be marked may be determined according to the difference between a set consisting of the embedding vectors of all first nodes determined in S108b4 and a set consisting of the first nodes in all subgraphs to be marked.

[0143] Exemplarily, the first node in the subgraph to be marked is a node of a determined type in the graph structure. The embedding vector of the first node includes all the first nodes in the graph structure. By taking the difference between the set corresponding to the embedding vector and the set corresponding to the node of the determined type, all nodes of undetermined type in the graph structure are obtained and determined as the first node to be marked.

[0144] In S108b6, the type of the first node to be marked is determined according to the clustering algorithm.

[0145] Exemplarily, the clustering algorithm may be a Kmeans algorithm.

[0146] In one example, a clustering algorithm can be used to determine each first node to be marked, thereby determining the type of the historical search results. For example, a type identifier can be marked for each first node to be marked to determine the type of each first node to be marked.

[0147] Optionally, after determining the type of the first node to be marked, a target marked sample may be determined according to the type of the first node obtained in S108b3 and S108b6. The target marked sample represents the marked result of the historical search result in the sample to be marked.

[0148] Exemplarily, the type of the first node in the subgraph to be labeled and the type of the first node to be labeled may be merged to obtain a target labeled sample.

[0149] In the above embodiment, the type of the first node with complex association relationship in the graph structure is determined by a clustering algorithm, so as to determine the type of the historical search results in the sample to be marked, thereby improving the accuracy and efficiency of determining the search result type.

[0150] In a possible implementation, user behavior data of each search result of the historical search term is aggregated to obtain a user satisfaction value of each historical search result, including: determining the user satisfaction of the historical search results based on the aggregated user behavior data; and determining the user satisfaction value of the historical search results based on the user satisfaction.

[0151] Optionally, determining user satisfaction of historical search results based on aggregated user behavior data includes: determining user satisfaction of historical search results based on each type of user behavior data of each user on historical search results and a weight coefficient of each type of user behavior data.

[0152] In the above embodiment, the satisfaction with the historical search results is determined by each user's behavior data on the historical search results and the weight coefficient of each user behavior data. The user behavior data of different user accounts can be used to characterize the user account's satisfaction with the historical search results under each search term.

[0153] Exemplarily, the user satisfaction of the search results satisfies the following formula:

[0154]

[0155] Among them, satisfy represents user satisfaction, X i Represents the i-th user behavior data, A i The weight coefficient that represents the i-th user behavior data, where n is a positive integer.

[0156] Optionally, the user behavior data includes at least one or more of the number of clicks on historical search results, the number of likes, and the number of follows on the author of the search results.

[0157] For example, satisfaction = 1*click+2*like+3*follow+…, where click represents the number of clicks on historical search results, like represents the number of likes on historical search results, follow represents the number of times the author of the historical search results is followed, … represents other user behavior data, 1 represents the weight coefficient of click, 2 represents the weight coefficient of like, and 3 represents the weight coefficient of follow.

[0158] Optionally, the weight coefficient of the user behavior data is predetermined according to the importance of the user behavior data.

[0159] In the above embodiment, the user satisfaction of the historical search results is determined by aggregating user behavior data, and the user satisfaction value of the historical search results is obtained according to the user satisfaction, thereby obtaining the historical search results that meet the user's search.

[0160] In one possible implementation, a user satisfaction value of historical search results is determined based on user satisfaction, including: determining a user satisfaction rate of the historical search results based on user satisfaction and the number of times the historical search results are played; determining a user satisfaction confidence of the historical search results based on the number of times the historical search results are played; and determining a user satisfaction value of the historical search results based on the user satisfaction rate and the user satisfaction confidence.

[0161] Exemplarily, the user satisfaction rate satisfies the following formula:

[0162]

[0163] Among them, satisfy_ratio represents the user satisfaction rate, satisfy represents the user satisfaction, play_cnt represents the number of times the search result is played, and N represents the normalization coefficient.

[0164] For example, when determining the value of N, the maximum value of the satisfaction count in a user behavior is taken. For example, in the above example, the maximum values ​​of click, like, and follow are all 1, and the corresponding weight coefficients of click, like, and follow are added, then N=6.

[0165] Optionally, play_cnt represents the number of times the search result is played, which is an aggregated value of the number of times the search result corresponds to the play count.

[0166] By determining the satisfaction rate of the search results in each sample set, it is possible to characterize the satisfaction rate values ​​of different users for a search result under each search.

[0167] Exemplarily, the user satisfaction confidence satisfies the following formula:

[0168]

[0169] Among them, satisfy confidence It represents the user satisfaction confidence, and play_cnt represents the number of times the search result is played.

[0170] It should be noted that the more times a search result is played, the higher the user confidence corresponding to the search result.

[0171] Exemplarily, the user satisfaction value satisfies the following formula:

[0172] satisfy score =satisfy_ratio*satisfy confidence

[0173] Among them, satisfy score Represents user satisfaction value, satisfaction confidence It represents the user satisfaction confidence, and satisfaction_ratio represents the user satisfaction rate.

[0174] Optionally, the user satisfaction value is used to represent the user's satisfaction with the search results.

[0175] In the above embodiment, the user satisfaction value of the historical search results is obtained by the number of times the user satisfaction of the historical search results is played, thereby determining the satisfaction of the user account with the search results, so as to facilitate screening of historical search results with high user satisfaction for type determination, that is, screening out historical search results that meet the search needs of the user account, and thus avoiding type determination of search results with low user satisfaction, that is, avoiding type determination of historical results that do not meet the search needs of the user account, thereby improving the effectiveness and efficiency of search structure type determination.

[0176] The above mainly introduces the solution provided by the embodiment of the present application from the perspective of the method. In order to realize the above functions, it includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of each example described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware 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 exceed the scope of the present application.

[0177] The embodiment of the present disclosure also provides a content search device.

[0178] Figure 6is a block diagram of a content search device according to an exemplary embodiment. Figure 6 The device includes an acquisition module 601, a reading module 602 and a display module 603.

[0179] The acquisition module 601 is configured to acquire a search term and determine a plurality of search results corresponding to the search term; for example, Figure 1 , the acquisition module 601 can be used to execute S101.

[0180] The reading module 602 is configured to read the pre-stored correspondence between the search term and at least two types of search results; the correspondence is determined based on the historical search results whose user satisfaction value and display times corresponding to the search term meet the preset conditions. Figure 1 , the reading module 602 can be used to execute S102.

[0181] The acquisition module 601 is further configured to acquire at least two types of target search results from the multiple search results according to the corresponding relationship. Figure 1 , the acquisition module 601 can be used to execute S103.

[0182] The display module 603 is configured to display the target search results. Figure 1 , the display module 603 can be used to execute S104.

[0183] In a possible implementation, the device further includes a configuration module configured to perform the following operations: obtaining historical search results corresponding to the search term within a preset time period, and user behavior data of the historical search results; aggregating the user behavior data of each historical search result of the search term to obtain a user satisfaction value of each historical search result; obtaining historical search results whose user satisfaction value is greater than a first threshold and whose display times are greater than a second threshold to obtain samples to be marked; determining the type of each historical search result in the sample to be marked to obtain a correspondence between the search term and at least two types of search results. For example, in combination with Figure 2 , the configuration module can be used to execute S105-S108.

[0184] In another possible implementation, the configuration module is specifically configured to execute: constructing a graph structure according to the sample to be labeled and the graph embedding model; the graph structure includes a first node for representing the historical search results in the sample to be labeled, a second node for representing the search term, and a connection relationship between the first node and the second node; determining the type of each historical search result in the sample to be labeled according to the graph structure. Figure 3, the configuration module can be used to execute S108a-S108b.

[0185] In another possible implementation, the configuration module is specifically configured to execute: determining a maximal connected subgraph of the graph structure; obtaining a maximal connected subgraph with a number of nodes less than a preset threshold to obtain a subgraph to be marked; the node includes the first node and the second node; and determining the type of the first node in the subgraph to be marked. Figure 4 , the configuration module can be used to execute S108b1-S108b3. In another possible implementation, the configuration module is specifically configured to execute: determining the embedding vector of the first node in the graph structure; determining the first node to be marked according to the embedding vector and the first node in the subgraph to be marked; and determining the type of the first node to be marked according to the clustering algorithm. For example, in combination Figure 5 , the configuration module can be used to execute S108b4-S108b6.

[0186] In another possible implementation, the configuration module is specifically configured to execute: determining user satisfaction of the historical search results based on the aggregated user behavior data; and determining a user satisfaction value of the historical search results based on the user satisfaction.

[0187] In another possible implementation, the configuration module is specifically configured to execute: determining user satisfaction with the historical search results based on each user behavior data of each user on the historical search results and a weight coefficient of each user behavior data.

[0188] In another possible implementation, the configuration module is specifically configured to perform: determining a user satisfaction rate of the historical search results based on the user satisfaction and the number of times the historical search results are played; determining a user satisfaction confidence level of the historical search results based on the number of times the historical search results are played; and determining a user satisfaction value of the historical search results based on the user satisfaction rate and the user satisfaction confidence level.

[0189] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0190] Figure 7 700 is a block diagram of an electronic device according to an exemplary embodiment. The electronic device 700 may be a mobile phone, a tablet computer, a laptop computer, a PDA, a desktop computer, a portable computer, a vehicle-mounted terminal, a wearable device, etc. Figure 7As shown, the electronic device 700 includes but is not limited to a processor 701, a memory 702, a display 703, an input unit 704, an output unit 705, a network unit 706, an interface unit 707, a radio frequency unit 708, a sensor 709, and a power supply 710.

[0191] It should be noted that those skilled in the art can understand that Figure 7 The structure of the electronic device 700 shown in the figure does not constitute a limitation on the electronic device 700, and the electronic device 700 may include Figure 7 More or fewer components may be shown, or certain components may be combined, or arranged differently.

[0192] In the embodiment of the present disclosure, the processor 701 is used to execute instructions in the memory to implement the following Figure 1-Figure 5 Use any of the content search methods shown in .

[0193] It should be noted that the above-mentioned electronic device 700 can implement each process implemented by the electronic device in the above-mentioned method embodiment, and can achieve the same technical effect. To avoid repetition, the detailed description will not be repeated here.

[0194] The processor 701 is the control center of the electronic device 700. It uses various interfaces and lines to connect various parts of the entire electronic device 700. By running or executing software programs and / or modules stored in the memory 702, and calling data stored in the memory 702, it executes various functions of the electronic device 700 and processes data, thereby monitoring the electronic device 700 as a whole. The processor 701 may include one or more processing units. In one embodiment, the processor 701 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user pages, and application programs, etc., and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 701.

[0195] The memory 702 can be used to store software programs and various data. The memory 702 can mainly include a program storage area and a data storage area, wherein the program storage area stores an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, phone book), etc. In addition, the memory 702 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0196] The display 703 is used to display the text special effects, that is, the display 703 can display the animation of the text special effects. The display 703 may include a display panel, and the display panel may be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), etc.

[0197] The input unit 704 may be used to receive audio or video signals. The input unit 704 may include a graphics processing unit (GPU) and a microphone, and the image processor processes the image data of a static picture or video obtained by an image capture device (such as a camera) in a video capture mode or an image capture mode. The processed image frame can be displayed on the display 703. The image frame processed by the image processor can be stored in the memory 702 (or other storage medium) or sent via the radio frequency unit 708 or the network unit 706. The microphone can receive sound and can process such sound into audio data. The processed audio data can be converted into a format output that can be sent to a mobile communication base station via the radio frequency unit 708 in the case of a telephone call mode.

[0198] The input unit 704 can input search information for the user, can be used to receive input digital or character information, and generate key signal input related to the user settings and function control of the electronic device 700. Specifically, the input unit includes a touch panel and other input devices. The touch panel, also known as a touch screen, can collect the user's touch operations on or near it (such as the user's operation on or near the touch panel using any suitable object or accessory such as a finger, stylus, etc.). The touch panel may include two components: a touch detection device and a touch controller. Among them, the touch detection device detects the user's touch direction, detects the signal brought by the touch operation, and transmits the signal to the touch controller; the touch controller receives the touch information from the touch detection device, and converts it into contact coordinates, and then sends it to the processor 701, receives the command sent by the processor 701 and executes it. In addition, the touch panel can be implemented using various types such as resistive, capacitive, infrared and surface acoustic wave. When processing the touch panel, the user input unit may also include other input devices. Specifically, other input devices may include but are not limited to a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a trackball, a mouse, and a joystick, which will not be repeated here.

[0199] Furthermore, the touch panel can be covered on the display panel. When the touch panel detects a touch operation on or near it, it is transmitted to the processor 701 to determine the type of touch time, and then the processor 701 provides corresponding data output on the display panel according to the type of touch event. The touch panel and the display panel can be used as two independent components to realize the input and output functions of the electronic device 700, or the touch panel and the display panel can be integrated to realize the input and output functions of the electronic device 700, which is not limited here.

[0200] The output unit 705 may be an audio output unit, which may convert the audio data received by the RF unit 708 or the network unit 706 or stored in the memory 702 into an audio signal and output the sound. Moreover, the audio output unit may also provide audio output related to a specific function performed by the electronic device 700 (e.g., a call signal reception sound, a message reception sound, etc.). The audio output unit includes a speaker, a buzzer, a receiver, etc.

[0201] The electronic device 700 provides the user with wireless broadband Internet access through the network unit 706, such as helping the user to send and receive emails, browse web pages, and access streaming media.

[0202] The interface unit 707 is an interface for connecting an external device to the electronic device 700. For example, the external device may include a wired or wireless headset port, an external power supply (or battery charger) port, a wired or wireless data port, a memory card port, a port for connecting a device with an identification module, an audio input / output (I / O) port, a video I / O port, a headphone port, etc. The interface unit 707 may be used to receive input (e.g., data information, power, etc.) from an external device and transmit the received input to one or more elements in the electronic device 700 or may be used to transmit data between the electronic device 700 and external devices.

[0203] The radio frequency unit 708 can be used to receive and send signals during the process of sending and receiving information or making calls. Specifically, after receiving the downlink data from the base station, it is sent to the processor 701 for processing; in addition, the uplink data is sent to the base station. Generally, the radio frequency unit 708 includes but is not limited to an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, etc. In addition, the radio frequency unit 708 can also communicate with the network and other devices through a wireless communication system.

[0204] Sensor 709 may include at least one of a light sensor, a motion sensor, and other sensors. Specifically, the light sensor includes an ambient light sensor and a proximity sensor, wherein the ambient light sensor can adjust the brightness of the display panel according to the brightness of the ambient light, and the proximity sensor can turn off the display panel and / or backlight when the electronic device 700 is moved to the ear. As a type of motion sensor, the accelerometer sensor can detect the magnitude of acceleration in each direction (generally three axes), and can detect the magnitude and direction of gravity when stationary, which can be used to identify the posture of the electronic device (such as horizontal screen switching, related games, magnetometer posture calibration), vibration recognition related functions (such as pedometer, tapping), etc.; sensor 709 may also include a fingerprint sensor, a pressure sensor, an iris sensor, a molecular sensor, a gyroscope, a barometer, a hygrometer, a thermometer, an infrared sensor, etc., which will not be repeated here.

[0205] The power supply 710 (such as a battery) can be used to power various components. In one embodiment, the power supply 710 can be logically connected to the processor 701 through a power management system, so that the power management system can manage charging, discharging, and power consumption.

[0206] In addition, the electronic device 700 also includes some functional modules (such as a camera) that are not shown, which will not be described here.

[0207] In one example, see Figure 6 The processing functions of the acquisition module 601, the reading module 602 and the display module 603 can all be performed by Figure 7 The processor 701 in the embodiment calls the computer program stored in the memory 702 for implementation.

[0208] In an exemplary embodiment, the present disclosure also provides a computer-readable storage medium including instructions, such as a memory 702 including instructions, which can be executed by a processor 701 of a server 700 to complete the Figures 1 to 5 The content search method shown in .

[0209] Optionally, the computer-readable storage medium may be a non-temporary computer-readable storage medium, for example, the non-temporary computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0210] In an exemplary embodiment, the present disclosure also provides a computer program product including one or more instructions, which can be executed by the processor 701 of the server 700 to complete Figures 1 to 5 The content search method shown in .

[0211] It should be noted that when the instructions in the above-mentioned computer-readable storage medium or one or more instructions in the computer program product are executed by the processor 701 of the server 700, each process of the above-mentioned content search method embodiment is implemented, and the above-mentioned content search method embodiment can be achieved. Figures 1 to 5 The content search method shown in has the same technical effect, so to avoid repetition, it will not be repeated here.

[0212] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.

[0213] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A content search method, used in an electronic device, characterized in that: include: Obtaining a search term and determining a plurality of search results corresponding to the search term; Reading a pre-stored correspondence between the search term and at least two types of search results; The corresponding relationship is determined based on the historical search results whose user satisfaction value and display times corresponding to the search term meet the preset conditions respectively; According to the corresponding relationship, obtaining at least two types of target search results from the multiple search results; Display the target search results; Obtaining historical search results corresponding to the search term within a preset time period, and user behavior data of the historical search results; Aggregating user behavior data of each historical search result of the search term to obtain a user satisfaction value of each historical search result; Obtain historical search results whose user satisfaction value is greater than a first threshold and whose display times are greater than a second threshold, and obtain samples to be marked; Constructing a graph structure according to the samples to be labeled and the graph embedding model; The graph structure includes a first node for representing the historical search results in the sample to be marked, a second node for representing the search term, and a connection relationship between the first node and the second node; According to the graph structure, the type of each historical search result in the sample to be marked is determined, and a corresponding relationship between the search term and at least two types of search results is obtained.

2. The content search method according to claim 1, characterized in that: Determining the type of each historical search result in the sample to be marked according to the graph structure includes: Determine a maximal connected subgraph of the graph structure; Acquire a maximal connected subgraph whose number of nodes is less than a preset threshold, and obtain a subgraph to be marked; the nodes include the first node and the second node; Determine the type of the first node in the subgraph to be marked.

3. The content search method according to claim 1, characterized in that: The step of determining the type of each historical search result in the sample to be marked according to the graph structure further includes: Determining an embedding vector of a first node in the graph structure; Determine a first node to be marked according to the embedding vector and the first node in the subgraph to be marked; According to a clustering algorithm, the type of the first node to be marked is determined.

4. The content search method according to any one of claims 1 to 3, characterized in that: The aggregating the user behavior data of each search result of the historical search term to obtain the user satisfaction value of each historical search result includes: Determining user satisfaction with the historical search results based on the aggregated user behavior data; A user satisfaction value of the historical search results is determined according to the user satisfaction.

5. The content search method according to claim 4, characterized in that: Determining the user satisfaction of the historical search results according to the aggregated user behavior data includes: The user satisfaction of the historical search results is determined according to each user behavior data of each user to the historical search results and a weight coefficient of the each user behavior data.

6. The content search method according to claim 4, characterized in that: Determining the user satisfaction value of the historical search results according to the user satisfaction includes: Determining a user satisfaction rate of the historical search results according to the user satisfaction and the number of times the historical search results are played; Determining the user satisfaction confidence of the historical search results according to the number of times the historical search results are played; A user satisfaction value of the historical search results is determined according to the user satisfaction rate and the user satisfaction confidence.

7. A content search device, characterized in that: include: An acquisition module, configured to acquire a search term and determine a plurality of search results corresponding to the search term; A reading module, configured to read a pre-stored correspondence between the search term and at least two types of search results; The corresponding relationship is determined based on the historical search results whose user satisfaction value and display times corresponding to the search term meet the preset conditions respectively; The acquisition module is further configured to acquire at least two types of target search results from the multiple search results according to the corresponding relationship; A display module, configured to display the target search results; A configuration module is configured to execute: obtaining historical search results corresponding to the search term within a preset time period, and user behavior data of the historical search results; Aggregating user behavior data of each historical search result of the search term to obtain a user satisfaction value of each historical search result; Obtain historical search results whose user satisfaction value is greater than a first threshold and whose display times are greater than a second threshold, and obtain samples to be marked; Constructing a graph structure according to the sample to be labeled and the graph embedding model; the graph structure includes a first node for representing the historical search results in the sample to be labeled, a second node for representing the search term, and a connection relationship between the first node and the second node; According to the graph structure, the type of each historical search result in the sample to be marked is determined, and a corresponding relationship between the search term and at least two types of search results is obtained.

8. The content search device according to claim 7, characterized in that: The configuration module is specifically configured to execute: Determine a maximal connected subgraph of the graph structure; Acquire a maximal connected subgraph whose number of nodes is less than a preset threshold, and obtain a subgraph to be marked; the nodes include the first node and the second node; Determine the type of the first node in the subgraph to be marked.

9. The content search device according to claim 8, characterized in that: The configuration module is specifically configured to execute: Determining an embedding vector of a first node in the graph structure; Determine a first node to be marked according to the embedding vector and the first node in the subgraph to be marked; According to a clustering algorithm, the type of the first node to be marked is determined.

10. The content search device according to any one of claims 7 to 9, characterized in that: The configuration module is specifically configured to execute: Determining user satisfaction with the historical search results based on the aggregated user behavior data; A user satisfaction value of the historical search results is determined according to the user satisfaction.

11. The content search device according to claim 10, characterized in that: The configuration module is specifically configured to execute: The user satisfaction of the historical search results is determined according to each user behavior data of each user to the historical search results and a weight coefficient of the each user behavior data.

12. The content search device according to claim 10, characterized in that: The configuration module is specifically configured to execute: Determining a user satisfaction rate of the historical search results according to the user satisfaction and the number of times the historical search results are played; Determining the user satisfaction confidence of the historical search results according to the number of times the historical search results are played; A user satisfaction value of the historical search results is determined according to the user satisfaction rate and the user satisfaction confidence.

13. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the content search method according to any one of claims 1 to 6.

14. A computer-readable storage medium, characterized in that: When the instructions in the computer-readable storage medium are executed by a processor of the server, the electronic device is enabled to execute the content search method according to any one of claims 1 to 6.

15. A computer program product, characterized in that The computer program product includes computer instructions, and when the computer instructions are executed on a server, the electronic device executes the content search method according to any one of claims 1 to 6.

Images