Data identification method and device, computer device, and storage medium

Abstract

Embodiments of the present application disclose a data recognition method and device, computer equipment and a storage medium. By using machine learning in an artificial intelligence model, object interest point pairs including object identification and content type identification are used to determine overall content completion parameters, single-type completion parameters and overall object completion parameters. The association between the overall content completion parameters representing the overall and the single-type completion parameters representing the single body, and the association between the overall object completion parameters representing the overall and the single-type completion parameters representing the single body are determined to determine whether the object interest point pairs are accurate. The completion degree of the object viewing the content can reflect the object preference to a certain extent. The three completion parameters determined by multiple levels and multiple dimensions are used to analyze the preference of the target object, which not only improves the accuracy of the object interest point pairs, but also improves the efficiency by analyzing the overall dimension, so as to quickly judge the accuracy of the object interest parameters based on more accurate object interest point pairs.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data processing, and in particular to a data identification method and device, computer equipment and a storage medium. BACKGROUND

[0002] A user interest portrait is the core of a content product and can identify the personalized preferences of different users for content. Through a user interest portrait, personalized content services such as content distribution and recommendation can be achieved. A user interest portrait can be quantified by a user interest point pair (pid-tag pair, where pid represents a user identifier and tag represents a content type identifier). If it is determined that a user a has an interest in content of type b, a corresponding user interest point pair can be established, where the pid in the pair is the user identifier of user a and the tag is type b. A user can have multiple interest points on content, which can be identified by multiple user interest point pairs.

[0003] Since the accuracy of a user interest portrait directly affects the effectiveness of subsequent content services, the accuracy of a user interest portrait needs to be reasonably evaluated. Currently, the accuracy evaluation of a user interest portrait is very time-consuming and prone to deviation. SUMMARY

[0004] To solve the above technical problems, the present application provides a data identification method, device, computer equipment and storage medium, which are used to solve the problem of time-consuming and low accuracy in evaluating a user interest portrait based on user research or in the form of a questionnaire.

[0005] The embodiments of the present application disclose the following technical solutions:

[0006] In one aspect, the present application provides a data identification method, which comprises:

[0007] Obtaining a user interest point pair, the user interest point pair comprising a user identifier and a content type identifier;

[0008] Determining an overall content completion parameter for different content types of content through the user identifier, and a single-type completion parameter for a target content corresponding to the content type identifier;

[0009] Based on the content type identifier, determining an overall user completion parameter for the target content by different users;

[0010] According to the association relationship between the overall content completion parameter and the single-type completion parameter, and the association relationship between the overall user completion parameter and the single-type completion parameter, determining whether the user interest point pair is accurate.

[0011] On the other hand, the present application provides a data identification device, the device comprising: an acquisition unit, a determination unit, and a judgment unit;

[0012] The acquiring unit is configured to acquire a user interest point pair, wherein the user interest point pair includes a user identifier and a content type identifier;

[0013] The determining unit is configured to determine, based on the user identifier, overall content completion parameters for contents of different content types, and single-type completion parameters for target content corresponding to the content type identifier;

[0014] determining overall user completion parameters for the target content by different users based on the content type identifier;

[0015] The judgment unit is used to determine whether the user interest point pair is accurate based on the association relationship between the overall content completion parameter and the single type completion parameter, and the association relationship between the overall user completion parameter and the single type completion parameter.

[0016] In another aspect, the present application provides a computer device, comprising a processor and a memory:

[0017] The memory is used to store program code and transmit the program code to the processor;

[0018] The processor is configured to execute the method described above according to the instructions in the program code.

[0019] On the other hand, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium is used to store a computer program, and the computer program is used to execute the method described in the above aspects.

[0020] In another aspect, embodiments of the present application provide a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method described in the above aspects.

[0021] It can be seen from the technical solution that the user interest point pair includes a user identifier and a content type identifier, and the overall content completion parameter, the single-type completion parameter and the overall user completion parameter can be determined. The overall content completion parameter represents the completion degree of each user viewing the corresponding different content types of content from the user level, the overall user completion parameter represents the completion degree of different users viewing each content type of content from the content type level, and the single-type completion parameter represents the completion degree of each user viewing each content type of content. The association between the overall content completion parameter representing the whole and the single-type completion parameter representing the single body, and the association between the overall user completion parameter representing the whole and the single-type completion parameter representing the single body can be determined to determine whether the user interest point pair is accurate. The completion degree of the user viewing the content can reflect the user's preference to a certain extent. Therefore, the multi-level and multi-dimensional analysis of the user's preference not only improves the accuracy of the user interest point pair, but also improves the efficiency by combining the overall dimension for analysis. Therefore, the accuracy of the user interest portrait can be quickly judged based on the more accurate user interest point pair, thereby improving the service effect of subsequent content. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0023] Figure 1 An application scenario diagram of a data recognition method provided by the embodiments of the present application;

[0024] Figure 2 A flowchart of a data recognition method provided by the embodiments of the present application;

[0025] Figure 3 A schematic diagram of a business logic using a user interest portrait provided by the embodiments of the present application;

[0026] Figure 4 A schematic diagram of a data recognition device provided by the embodiments of the present application;

[0027] Figure 5 A structural schematic diagram of a server provided by the embodiments of the present application;

[0028] Figure 6 A structural schematic diagram of a terminal device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0029] The embodiments of the present application will be described below with reference to the accompanying drawings.

[0030] In evaluating the user interest portrait, the related art mainly judges the user interest portrait by means of user research or questionnaire, but both the user research and the questionnaire are case by case, and the case by case manner is more inclined to qualitative analysis, and in the later stage, quantitative analysis may exist problems such as lack of confidence caused by too small sample size or result deviation caused by sample range selection deviation.

[0031] Based on this, the present application provides a data recognition method and device, computer equipment and storage medium, through multi-level and multi-dimensional analysis, determine whether the user interest point pair used to quantify the user interest portrait is accurate, so as to improve the service effect of subsequent content based on more accurate user interest portrait.

[0032] The data recognition method provided by the embodiments of the present application is realized based on artificial intelligence. Artificial intelligence (AI) is to use digital computers or digital computer controlled machines to simulate, extend and expand human intelligence, perceive environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology of computer science, which tries to understand the essence of intelligence and produce a new intelligent machine that can react in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines, so that machines have the functions of perception, reasoning and decision-making.

[0033] Artificial intelligence technology is a comprehensive discipline, involving a wide range of fields, both hardware and software technologies. Artificial intelligence basic technologies generally include sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction system, mechatronics and other technologies. Artificial intelligence software technology mainly includes computer vision technology, speech processing technology, natural language processing technology and machine learning / deep learning and other directions.

[0034] In the embodiments of the present application, the artificial intelligence technology mainly involved includes the above machine learning / deep learning and other directions.

[0035] The data recognition method provided by the application can be applied to a data recognition device with data processing capability, such as a terminal device and a server. The terminal device can be a smart phone, a desktop computer, a notebook computer, a tablet computer, a smart speaker, a smart watch, and the like, but is not limited thereto. The server can be a stand-alone physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. The terminal device and the server can be directly or indirectly connected through wired or wireless communication, which is not limited in the application.

[0036] The data recognition device can have machine learning capability. Machine learning is a multi-disciplinary subject, involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory, and other disciplines. It is a specialized study of how computers simulate or implement human learning behavior to acquire new knowledge or skills, reorganize existing knowledge structure to continuously improve performance. Machine learning is the core of artificial intelligence and the fundamental approach to making computers intelligent, and its applications are widespread in various fields of artificial intelligence. Machine learning and deep learning usually include artificial neural networks and other technologies.

[0037] In the data recognition method provided by the embodiment of the application, the artificial intelligence model used mainly involves the application of machine learning, and the exposure strategy of the target content is regulated through game theory in machine learning.

[0038] In addition, the data recognition device provided by the embodiment of the application also has cloud computing capability, which determines whether the user interest point is accurate by using cloud computing capability through big data processing, so as to determine a more accurate user interest portrait. Big data refers to a collection of data that cannot be captured, managed and processed within a certain time range by conventional software tools, and is a massive, high-growth and diversified information asset that requires new processing modes to have stronger decision-making, insight discovery and process optimization capabilities. With the advent of the cloud era, big data has attracted more and more attention. Big data requires special technology to effectively process large amounts of data over time. Technologies suitable for big data include large-scale parallel processing databases, data mining, distributed file systems, distributed databases, cloud computing platforms, the Internet and scalable storage systems.

[0039] In order to facilitate understanding of the technical solutions of the application, the data recognition method provided by the embodiment of the application is introduced below in combination with an actual application scenario, taking a server as a data recognition device.

[0040] Referring to Figure 1 , Figure 1 is a schematic diagram of an application scenario of the data recognition method provided by the embodiment of the application. InFigure 1 In the application scenario shown, a server 100 is configured to determine whether a user interest point pair is accurate.

[0041] The server 100 acquires a user interest point pair, which can be used to quantify a user interest profile. The user interest point pair includes a user identifier and a content type identifier. Figure 1 In the application scenario shown, two users view two videos of different content types respectively. User A and user B correspond to user identifiers pidA and pidB respectively. The video of content type a (hereinafter referred to as video a) and the video of content type b (hereinafter referred to as video b) correspond to content type identifiers taga and tagb respectively. The server 100 acquires four user interest point pairs, namely, the pidA-taga pair, the pidA-tagb pair, the pidB-taga pair, and the pidB-tagb pair.

[0042] The server 100 can determine an overall content completion parameter, a single-type completion parameter, and an overall user completion parameter by acquiring the user identifier and the content type identifier included in the user interest point pair. Figure 1 In the application scenario shown, the average play completion of video a viewed by user A is 90%, the average play completion of video b viewed by user A is 20%, the average play completion of video a viewed by user B is 50%, and the average play completion of video b viewed by user B is 80%.

[0043] The overall content completion parameter represents the completion degree of each user viewing the content of different content types corresponding to the user. For example, for user A, the average play completion of video a and video b is (90%+20%) / 2=55%, i.e., the overall content completion parameter of user A is 55%. Similarly, the overall content completion parameter of user B is (50%+80%) / 2=65%.

[0044] The overall user completion parameter represents the completion degree of different users viewing the content of each content type. For example, for video a, the average play completion of video a viewed by user A and user B is (90%+50%) / 2=70%, i.e., the overall user completion parameter of video a is 70%. Similarly, the overall user completion parameter of video b is (20%+80%) / 2=50%.

[0045] The single-type completion parameter represents the completion degree of each user viewing the content of each content type. For example, the single-type completion parameter of user A for video a is 90%, the single-type completion parameter of user A for video b is 20%, the single-type completion parameter of user B for video a is 50%, and the single-type completion parameter of user B for video b is 80%.

[0046] The completion degree of the user viewing the content can reflect the user's preference to some extent. By associating the overall content completion parameter representing the overall whole with the single type completion parameter representing the single body, and associating the overall user completion parameter representing the overall whole with the single type completion parameter, it can be determined whether the user interest point pair is accurate.

[0047] For example, if the overall content completion parameter is greater than the single type completion parameter, and the overall user completion parameter is greater than the single type completion parameter, it is determined that the user interest point pair is not accurate, for example, the pidA-tagb pair and the pidB-taga pair are both not accurate, user A is not interested in video b, and user B is not interested in video a. If the overall content completion parameter is less than the single type completion parameter, and the overall user completion parameter is less than the single type completion parameter, it is determined that the user interest point pair is accurate, for example, the pidA-taga pair and the pidB-tagb pair are both accurate, user A is interested in video a, and user B is interested in video b.

[0048] Therefore, by analyzing the user's preference from multiple aspects and multiple dimensions, not only the accuracy of the user interest point pair is improved, but also the efficiency is improved by analyzing in combination with the overall dimension. Further, based on the more accurate user interest point pair, it can be quickly judged whether the user interest portrait is accurate or not, thereby improving the service effect of subsequent content.

[0049] Next, with reference to the accompanying drawings, a data recognition method provided by an embodiment of the present application is described, taking a server as a data recognition device.

[0050] Referring to Figure 2 , Figure 2 A flowchart of a data recognition method provided by an embodiment of the present application is shown in FIG. 1. As shown in FIG. 1, the data recognition method includes the following steps: Figure 2

[0051] S201: Obtain a user interest point pair.

[0052] The user interest portrait can be quantified by a user interest point pair (pid-tag pair), so that the problem of judging the user interest portrait can be converted into determining whether the user interest point pair is accurate or not. In order to determine whether the user interest point pair is accurate, the user interest point pair can be obtained first, wherein the user interest point pair includes a user identifier and a content type identifier, the user identifier is used to identify a user, and the content type identifier is used to identify the type corresponding to the content.

[0053] The content is not specifically limited in the embodiment of the present application, for example, the content can be video content, song content, text content, etc. Different contents can be refined into different content types according to requirements, for example, the content type of the video content can be comedy type, action type, love type, etc.​

[0054] The embodiment of the present application does not specifically limit the acquisition manner of the user interest point pair, and the following is described by taking one manner as an example.

[0055] The N contents corresponding to the content type identifier are distributed to the M users, and the user interest point pair is determined according to the content completion condition of the distributed contents by the M users. Wherein, N and M are integers greater than or equal to 1.

[0056] For example, N contents belonging to one content type are distributed to M users, and the N contents are viewed X times. Taking an example that the M users view the N contents, X = N*M. According to the content completion condition of the contents viewed X times, for example, the user identifier corresponding to the user whose content completion condition is greater than a preset threshold and the content type identifier corresponding to the content are selected to form the user interest point pair. For another example, the users are sorted according to the content completion condition, and the user identifier corresponding to the top Z users and the content type identifier are selected to form the user interest point pair.

[0057] S202: Determine the overall content completion parameter of the content of different content types through the user identifier, and the single-type completion parameter of the target content corresponding to the content type identifier.

[0058] The user identifier can be used to determine the user. Taking one user as an example, the single-type completion parameter of the target content corresponding to the content type identifier of the user is determined. For example, the single-type completion parameter is used to identify the average play completion condition of the content corresponding to the content type identifier through the user identifier. Taking a video as an example, the average play completion condition of each user in each content type of video viewed by the user.

[0059] For example, user A only views videos of content types X and Y, wherein the play completion condition of the video of content type X viewed by user A is 90%, and the play completion condition of the video of content type Y viewed by user A is 100%, the single-type completion parameter of user A for video X is 90%, and the single-type completion parameter of user A for video Y is 100%.

[0060] In order to improve the accuracy of determining the user interest point pair, the analysis is carried out from multiple dimensions, that is, not only the single-type completion parameter representing the single body is determined, but also the content completion parameter representing the whole is determined.

[0061] Still taking one user as an example, the overall content completion parameter of the content of all content types viewed by the user is determined, so as to determine the completion degree of each user in the content of different content types corresponding to the user from the user level. For example, the overall content completion parameter is used to identify the average play completion condition of the content of different content types through the user identifier. Taking a video as an example, the average play completion condition of each user in all content types of video viewed by the user.

[0062] For example, user A only views videos of content types X and Y, where video x1 and video x2 belong to videos of content type X, user A views video x1 and the playing completion of video x1 is 90%, and user A views video x2 and the playing completion of video x2 is 100%, then the average playing completion of videos of content type X viewed by user A is 95%, and the average playing completion of videos of content type Y viewed by user A is 85%, and the overall content completion parameter of user A is 90%.

[0063] S203: determining an overall user completion parameter of the target content by different users based on the content type identifier.

[0064] In order to improve the accuracy of determining the user interest point pair, analysis is performed from multiple aspects, that is, not only the overall content completion parameter is determined from the user aspect, but also the overall user completion parameter is determined from the content type aspect.

[0065] The content type identifier can be used to determine the content type of the target content, and the overall user completion parameter of all users viewing the target content corresponding to the content type is determined for one content type. For example, the overall user completion parameter is used to identify the average playing completion of different users to the content corresponding to the content type identifier, and for videos, the average playing completion of all users viewing videos of one content type.

[0066] For example, user A and user B both view videos of content type X, where user A views videos of content type X and the playing completion is 90%, and user B views videos of content type X and the playing completion is 80%, and the overall user completion parameter of content type X is 85%.

[0067] It should be noted that the present application does not specifically limit the order of execution of S202 and S203, for example, S202 and S203 can be executed simultaneously.

[0068] S204: determining whether the user interest point pair is accurate according to the association relationship between the overall content completion parameter and the single-type completion parameter, and the association relationship between the overall user completion parameter and the single-type completion parameter.

[0069] The completion degree of content viewed by a user can reflect the preference of the user to a certain extent, and thus the overall content completion parameter, the single-type completion parameter and the overall user completion parameter determined by multiple aspects and multiple dimensions can analyze the preference of the user, improve the accuracy of analysis, and compared with the single-point analysis in the related art, the overall analysis also improves the efficiency of analysis, so as to quickly determine whether the user interest point pair representing the preference of the user is accurate.

[0070] According to the association between the overall content completion parameter and the single-type completion parameter, that is, by analyzing the completion degree of the target content corresponding to one content type viewed by a user and the average overall completion degree of all content types viewed by the user, the preference difference of the user for the target content and other content can be understood for the user himself, so as to determine the preference degree of the user for the target content.

[0071] According to the association between the overall user completion parameter and the single-type completion parameter, that is, by analyzing the completion degree of the target content corresponding to one content type viewed by a user and the average overall completion degree of all users for the target content, the preference difference of the user for the target content and other users for the target content can be understood for the target content, so as to determine the preference degree of the user for the target content.

[0072] The embodiments of the present application do not specifically limit the way of determining whether the user interest point pair is accurate, and the following will be described by taking the comparison of the size relationship as an example.

[0073] If the overall content completion parameter is greater than the single-type completion parameter, and the overall user completion parameter is greater than the single-type completion parameter, that is, by longitudinal analysis, it is determined that the preference of a user for the target content is low, and by transverse analysis, it is determined that the preference of the user for the target content is low compared to all content types viewed by the user, which indicates that the user is not interested in the target content, that is, the user interest point pair composed of the user identifier corresponding to the user and the content type identifier corresponding to the target content is inaccurate.

[0074] Similarly, if the overall content completion parameter is less than or equal to the single-type completion parameter, and the overall user completion parameter is less than or equal to the single-type completion parameter, that is, by longitudinal analysis, it is determined that the preference of a user for the target content is high, and by transverse analysis, it is determined that the preference of the user for the target content is high compared to all content types viewed by the user, which indicates that the user is interested in the target content, that is, the user interest point pair composed of the user identifier corresponding to the user and the content type identifier corresponding to the target content is accurate.

[0075] Therefore, by transverse analysis and longitudinal analysis, it can be determined whether the user interest point pair is accurate, and then the accuracy of the user interest portrait can be quickly judged based on the more accurate user interest point pair, so as to improve the service effect of subsequent content.

[0076] In order to obtain a more accurate user interest portrait, after determining whether the user interest point pair is accurate, the user interest portrait is adjusted based on the accuracy result of whether the user interest point pair is accurate.

[0077] For example, the inaccurate user interest point pair in the user interest portrait is deleted to obtain a modified user interest portrait. For another example, the weight of the accurate user interest point pair in the user interest portrait is increased, and the weight of the inaccurate user interest point pair in the user interest portrait is decreased to obtain a modified user interest portrait, which is more accurate than the user interest portrait before modification.

[0078] The user interest portrait adjusted by the accuracy result of the user interest point pair can more accurately reflect the personalized preferences of the corresponding user, thereby better providing content services for the user.

[0079] In a possible implementation, a user interest portrait of a user to be recommended is obtained, and the user interest point pair in the user interest portrait is adjusted according to the corresponding accuracy result;

[0080] The association degree between the user to be recommended and the content to be recommended is determined according to the user interest portrait of the user to be recommended, and the target content in the content to be recommended that meets a preset condition is recommended to the user to be recommended. For example, the content to be recommended that has an association degree greater than a preset threshold (such as 90%, 95%, etc.) is recommended as the target content and recommended to the user to be recommended. For another example, the content to be recommended is sorted according to the association degree, and the first N (such as the first 10, the first 3, etc.) content to be recommended is recommended as the target content and recommended to the user to be recommended.

[0081] The preset threshold can be in a range of 50%-100%, and the value of N can be a value greater than or equal to 1, which is not limited in the present application.

[0082] That is, when a user needs to be recommended content, the user interest portrait of the user is determined, which can be obtained by adjusting the above scheme. Since the user interest portrait is more in line with the actual preferences of the user than before adjustment, the content determined according to the user interest portrait is more in line with the needs of the user when the content is recommended to the user, thereby significantly improving the success rate of content recommendation.

[0083] In addition, the adjusted user interest portrait can be applied to various content-related scenarios or scenarios requiring user identification and clustering based on the user interest portrait. As can be seen from the above technical solutions, the user interest point pair includes a user identifier and a content type identifier, and the overall content completion parameter, the single-type completion parameter, and the overall user completion parameter can be determined. The overall content completion parameter represents the completion degree of each user viewing the corresponding different content types of content from the user level, the overall user completion parameter represents the completion degree of different users viewing each content type of content from the content type level, and the single-type completion parameter represents the completion degree of each user viewing each content type of content. By associating the overall content completion parameter representing the overall and the single-type completion parameter representing the single body, and the overall user completion parameter representing the overall and the single-type completion parameter representing the single body, it can be determined whether the user interest point pair is accurate. The completion degree of the user viewing the content can reflect the user's preference to some extent. Therefore, by analyzing the user's preference from multiple levels and multiple dimensions, the accuracy of the user interest point pair is improved, and the analysis combined with the overall dimension also improves the efficiency. Therefore, based on the more accurate user interest point pair, the accuracy of the user interest portrait can be quickly judged, thereby improving the service effect of subsequent content.

[0084] As a possible implementation, after obtaining the accuracy result for identifying whether the user interest point pair is accurate, a vector space can be constructed based on the accuracy result of the user interest point pair, so that the accuracy of other user interest point pairs can be quickly determined based on the constructed vector space. Specifically, a vectorized representation of the user interest point pair is determined, and a vector space is constructed according to the vectorized representation and the accuracy result of the user interest point pair.

[0085] For example, the accuracy result of the user interest point pair can obtain the proportion of accurate or inaccurate user interest point pairs (accurate or inaccurate is represented by 0 or 1). Then, the accuracy of a single user interest point pair is calculated according to the probability distribution, and a two-dimensional vector distribution, i.e., a vector space, is obtained, wherein the vector space can take the user identifier and the content type identifier as the coordinate axis to form a space grid. Referring to Table 1, which is a table for indicating whether the user interest point pair is accurate, wherein N represents inaccurate and Y represents accurate.

[0086] Table 1

[0087]

[0088] If the to-be-identified user interest point pair is obtained, the accuracy of the to-be-identified user interest point pair is determined according to the distribution relationship between the position of the to-be-identified user interest point pair in the vector space constructed in the foregoing and the position of the user interest point pair. For example, the to-be-identified user interest point pair is closer to the user interest point pair with a determined result of accuracy in the vector space, and the accuracy result of the to-be-identified user interest point pair is more accurate, so that different levels of measurement results can be counted

[0089] The present application does not specifically limit the measurement results of different dimensions, and the following three are taken as examples for description.

[0090] (1) User interest point pair level.

[0091] The proportion of the accuracy of the user interest point pair being greater than x% is counted, that is, the number of user interest point pairs with an accuracy greater than x% divided by the number of all user interest point pairs.

[0092] (2) User level.

[0093] For a user, the user can view contents of multiple content types, and the proportion of users with an accuracy greater than x% for one content type can be determined, and the proportion of users with an accuracy greater than x% for two content types can be determined.

[0094] (3) Content type level.

[0095] For a content type, the proportion of content types with an accuracy greater than y% for x% of users.

[0096] As a possible implementation manner, for a content-type product, the user interest portrait is an explicit factor for content personalization recommendation, which is used to represent the preference degree of each user for different content types. The user interest portrait generally includes multiple user interest point pairs. After the accuracy result of whether the to-be-identified user interest point pair is accurate is obtained, the weight of the to-be-identified user interest point pair in the user interest portrait is adjusted according to the accuracy of the to-be-identified user interest point pair, for example, the weight of the user interest point pair with high accuracy is increased, so as to increase the credibility of the user interest portrait, so that the user interest portrait is more accurate, and the user interest portrait recall is more accurate, and the next-day retention rate of the content-type product is effectively improved.

[0097] The measurement indicators of the user interest portrait can be divided into user interest portrait offline evaluation indicators and actual online user interest portrait application indicators. Among them, the user interest portrait offline evaluation indicators include coverage (user coverage, portrait coverage) and accuracy (user interest point pair accuracy, user accuracy, content type accuracy). The user interest portrait application indicators include application coverage and application effect. The application coverage is the exposure of the user interest portrait recall path / exposure ratio. The application effect can include a primary indicator - user interest portrait recall user consumption effect (complete playback / fast drawing); a secondary indicator - user interest portrait recall path exposure, duration, and next-day retention rate of affected users; and a tertiary indicator - user interest portrait recall path exposure, duration, and next-day retention rate of large plate users.

[0098] The data recognition method provided by the embodiments of the present application determines high-accuracy user interest point pairs and low-accuracy user interest point pairs in different application scenarios, and has different degrees of improvement on large plate indicators after application. Among them, the verification of the next-day retention rate improvement is relatively stable, about 0.4pp.

[0099] The present application does not specifically limit the application scenario, and the following will be described by taking five scenarios as examples.

[0100] (1) In-app recommendation

[0101] The user interest portrait is an independent recall path, including old and new user acceptance.

[0102] (2) In-app message stream push (push)

[0103] (3) Pre-scene recommendation for external triggering

[0104] Cooperate with other products to insert end content. Including chat software push, waterfall recommendation, news plug-in push, inner-layer recommendation, etc.

[0105] (4) User interest portrait special

[0106] For example, a user interest portrait for a drama user and a user interest portrait for a game user, so as to accept recommended content for special scenarios.

[0107] (5) Intelligent number package

[0108] Given the content, find the users suitable for consuming such content. It is usually used for content product operation support means, such as guaranteed volume mechanism, new hot distribution, etc.

[0109] The following will be described in combination with Figure 3 , taking three application scenarios as examples to describe the business logic using the user interest portrait. Referring to Figure 3 , the figure is a schematic diagram of a business logic using a user interest portrait provided by an embodiment of the present application.

[0110] (1) User clicks the page pushed by the out-of-end message stream, jumps to the out-of-end recommended page, if not an old user, downloads and activates to become a new user, and the new user logs in again to become an old user. If it is an old user, it is directly transferred to jump to the old user recommended page.

[0111] The out-of-end message stream pushing and the out-of-end recommendation are both based on the old user portrait, the user interest portrait with calculation factors, and the preset portrait for recommendation. The new user recommendation is based on the user interest portrait with calculation factors and the preset portrait for recommendation. The old user recommendation is based on the old user portrait for recommendation.

[0112] Among them, the old user portrait is a user interest portrait corresponding to a user with historical consumption behavior, the user interest portrait with calculation factors is a user interest portrait with historical consumption behavior and weights set for different content types, and the preset portrait is a user interest portrait without historical behavior but can be predicted.

[0113] (2) User clicks the page put by the external channel, downloads and activates to become a new user, and the new user logs in again to become an old user.

[0114] The external channel puts based on the old user portrait, the user interest portrait with calculation factors, and the preset portrait for recommendation.

[0115] (3) User jumps to the old user recommended page through the end-in message pushing.

[0116] The end-in message pushing is based on the old user portrait for recommendation.

[0117] As a possible implementation manner, in order to make the target content more accurate, the exposure strategy of the target content can also be adjusted in combination with the preference degree parameter and the exposure degree parameter. The ways of determining the preference degree parameter and the exposure degree parameter are introduced below.

[0118] (1) Preference degree parameter

[0119] The first preference degree mean value of different users to the target content and the second preference degree mean value of the same user to different content types are determined. For example, if there are m content types and n users, the first preference degree mean value can be represented as , and the second preference degree mean value can be represented as .

[0120] Among them, t represents the high consumption rate, u represents the user, and i represents the content type.

[0121] The preference degree parameter is determined according to the first preference degree mean value and the second preference degree mean value, and the application embodiment does not specifically limit the determination manner of the preference degree parameter, which is illustrated below by taking two ways as examples. Way one refers to formula (1), and way two refers to formula (2).

[0122] Method one:

[0123] (1)

[0124] Method two:

[0125] (2)

[0126] By combining the first preference mean and the second preference mean, the preference parameter is adjusted from multiple aspects, so that the preference parameter is more accurate.

[0127] (2) Exposure parameter

[0128] The first exposure mean of different users to the target content and the second exposure mean of the same user to different content types of content are determined. For example, if there are m content types and n users, the first exposure mean can be represented as , and the second exposure mean can be represented as .

[0129] Wherein, x represents the exposure proportion, u represents the user, and i represents the content type.

[0130] The exposure parameter is determined according to the first exposure mean and the second exposure mean, and the application embodiment does not specifically limit the determination method of the exposure parameter. The following two methods are taken as examples for description. Method one refers to formula (3), and method two refers to formula (4).

[0131] Method one:

[0132] (3)

[0133] Method two:

[0134] (4)

[0135] By combining the first exposure mean and the second exposure mean, the exposure parameter is adjusted from multiple aspects, so that the exposure parameter is more accurate.

[0136] After obtaining the preference parameter and the exposure parameter, the target content that the user likes but has little exposure and the target content that the user does not like but has much exposure can be determined according to the exposure parameter and the preference parameter, so as to regulate the exposure strategy of the target content.

[0137] For the target content to be regulated, the preference parameter and the exposure parameter of the target content can be obtained according to the above manner. If the preference parameter of the target content is greater than the preference parameter threshold, and the exposure parameter is less than the exposure parameter threshold, it indicates that the target content is more liked by the user but has less exposure in the user group. In this case, the exposure strategy corresponding to the target content can be adjusted to increase the number of recommendations of the target content in the user group.

[0138] For example, if the preference parameter of the target content is less than or equal to the preference parameter threshold, and the exposure parameter is greater than or equal to the exposure parameter threshold, it indicates that the target content is not liked by the user but has more exposure in the user group. The exposure strategy corresponding to the target content can be adjusted to reduce the number of recommendations of the target content in the user group, or even change the user group targeted by the target content.

[0139] The preference parameter threshold can be 50%-100%, and the exposure parameter threshold can be 50%-100%. The present application does not make specific limitations.

[0140] For example, when the preference parameter threshold is 80% and the exposure parameter threshold is 50%, if the preference parameter of the target content is 95% and the exposure parameter is 20%, the exposure strategy corresponding to the target content can be adjusted to increase the number of recommendations of the target content in the user group, for example, from the current 10 times to 50 times. If the preference parameter of the target content is 60% and the exposure parameter is 60%, the exposure strategy corresponding to the target content can be adjusted to reduce the number of recommendations of the target content in the user group, for example, from the current 30 times to 10 times. Next, the data recognition method provided by the embodiments of the present application will be described by taking the user interest portrait in the video application as an example.

[0141] By distributing N content corresponding to different content types to M users, the user interest point pair is determined according to the content completion of the distributed content by M users.

[0142] Then, the user interest point pair is evaluated, and the accuracy of the user interest point pair is determined by using the modes corresponding to S201-S204 to obtain the proportion of accurate and inaccurate user interest point pairs, and the accuracy of a single user interest point pair is calculated according to the probability distribution.

[0143] By longitudinal analysis, that is, comparing the consumption performance-completion rate of different users under the same content type, the content type is the user interest portrait of the user and the content type is not the user interest portrait of the user, and by transverse analysis, that is, comparing the consumption performance-completion rate of all videos under the same user, the video is the video corresponding to the user interest portrait of the user and the video is not the video corresponding to the user interest portrait of the user, a two-dimensional vector distribution is obtained, and a data model for evaluating user interest point pairs is constructed based on the vectorization of the user interest point pair and the accuracy result of the user interest point pair.

[0144] The to-be-identified user interest point pair is input into the data model, and the accuracy of the user interest point pair is obtained according to the position of the distribution of the completion rate in the vector space.

[0145] According to the accuracy of the to-be-identified user interest point pair, the weight of the to-be-identified user interest point pair in the user interest portrait is adjusted, so that the data model is more accurate, the user interest portrait recall is more accurate, and the next day retention rate of the user in the video application is improved.

[0146] The preference parameter and the exposure parameter distribution of the consumption content of different content types of different users in the video application are compared, so as to determine the content that the user likes but has little exposure, and the content that the user does not like but has much exposure, guide the directional content regulation of the business in the video application, so that the user is more willing to use the video application, and the next day retention rate of the user is improved.

[0147] For the data recognition method provided in the above embodiment, an embodiment of the present application further provides a data recognition device.

[0148] Referring to Figure 4 , the figure is a schematic diagram of a data recognition device provided by an embodiment of the present application. As Figure 4 shown, the data recognition device 400 comprises an acquisition unit 401, a determination unit 402 and a judgment unit 403;

[0149] The acquisition unit 401 is configured to acquire a user interest point pair, wherein the user interest point pair comprises a user identifier and a content type identifier;

[0150] The determination unit 402 is configured to determine an overall content completion parameter of content of different content types through the user identifier, and a single-type completion parameter of a target content corresponding to the content type identifier;

[0151] Based on the content type identifier, an overall user completion parameter of the target content by different users is determined;

[0152] The judgment unit 403 is used to determine whether the user interest point pair is accurate based on the association relationship between the overall content completion parameter and the single type completion parameter, and the association relationship between the overall user completion parameter and the single type completion parameter.

[0153] As a possible implementation manner, the determining unit 403 is configured to:

[0154] If the overall content completion parameter is greater than the single type completion parameter, and the overall user completion parameter is greater than the single type completion parameter, it is determined that the user interest point pair is inaccurate;

[0155] If the overall content completion parameter is less than or equal to the single-type completion parameter, and the overall user completion parameter is less than or equal to the single-type completion parameter, it is determined that the user interest point pair is accurate.

[0156] As a possible implementation, the acquiring unit 401 is configured to:

[0157] Distributing N contents corresponding to the content type identifier to M users, where N and M are integers greater than or equal to 1;

[0158] The user interest point pairs are determined according to the content completion status of the M users on the distributed content.

[0159] As a possible implementation manner, after obtaining the accuracy result for identifying whether the user interest point pair is accurate, the apparatus 400 is further configured to:

[0160] Determine the vectorized representation of the user's interest point pairs;

[0161] constructing a vector space according to the vectorized representation and the accuracy results of the user interest point pairs;

[0162] Obtain a pair of user interest points to be identified, and determine the accuracy of the pair of user interest points to be identified based on a distribution relationship between the pair of user interest points to be identified and the user interest point pairs in a vector space, where the vector space is constructed based on a vectorized representation of the user interest point pairs and an accuracy result of the user interest point pairs.

[0163] As a possible implementation, the apparatus 400 is further configured to:

[0164] According to the accuracy of the to-be-identified user interest point pair, the weight of the to-be-identified user interest point pair in the interest portrait of the user to which it belongs is adjusted.

[0165] As a possible implementation manner, the overall content completion parameter is used to identify average play completion of contents of different content types by the user identifier, the single-type completion parameter is used to identify average play completion of the content corresponding to the content type identifier by the user identifier, and the overall user completion parameter is used to identify average play completion of the content corresponding to the content type identifier by different users.

[0166] As a possible implementation manner, the apparatus 400 is further used to:

[0167] determine a first preference degree mean value of different users for the target content and a second preference degree mean value of the same user for contents of different content types;

[0168] determine the preference degree parameter according to the first preference degree mean value and the second preference degree mean value;

[0169] determine a first exposure degree mean value of different users for the target content and a second exposure degree mean value of the same user for contents of different content types;

[0170] determine the exposure degree parameter according to the first exposure degree mean value and the second exposure degree mean value;

[0171] regulate an exposure strategy of the target content according to the exposure degree parameter and the preference degree parameter.

[0172] The data recognition apparatus provided by the embodiment of the present application can determine the overall content completion parameter, the single-type completion parameter and the overall user completion parameter through the user interest point pair including the user identifier and the content type identifier. The overall content completion parameter represents the completion degree of each user viewing the corresponding contents of different content types from the user level, the overall user completion parameter represents the completion degree of different users viewing the contents of each content type from the content type level, and the single-type completion parameter represents the completion degree of each user viewing the contents of each content type. The association between the overall content completion parameter representing the whole and the single-type completion parameter representing the single body and the association between the overall user completion parameter representing the whole and the single-type completion parameter representing the single body can determine whether the user interest point pair is accurate. The completion degree of the user viewing the content can reflect the preference of the user to a certain extent. Therefore, the multi-level and multi-dimensional analysis of the preference of the user not only improves the accuracy of the user interest point pair, but also improves the efficiency by combining the overall dimension for analysis. Then, the accuracy of the user interest portrait can be quickly judged based on the more accurate user interest point pair, so as to improve the service effect of the subsequent content.

[0173] It should be noted that the information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0174] The embodiment of the present application also provides a computer device, which will be introduced from the perspective of hardware materialization.

[0175] Referring to Figure 5 , Figure 5 is a server structure schematic diagram provided by the embodiment of the present application. The server 1400 can have great difference due to different configurations or performances, and can include one or more central processing units (CPU) 1422 (for example, one or more processors) and a memory 1432, one or more storage media 1430 (for example, one or more mass storage devices) storing application programs 1442 or data 1444. The memory 1432 and the storage medium 1430 can be temporary storage or persistent storage. The programs stored in the storage medium 1430 can include one or more modules (not shown in the figure), and each module can include a series of instruction operations in the server. Further, the central processing unit 1422 can be configured to communicate with the storage medium 1430 and execute a series of instruction operations in the storage medium 1430 on the server 1400.

[0176] The server 1400 can also include one or more power supplies 1426, one or more wired or wireless network interfaces 1450, one or more input and output interfaces 1458, and / or one or more operating systems 1441, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

[0177] The steps performed by the server in the above embodiment can be based on the Figure 5 server structure shown in the figure.

[0178] The CPU 1422 is configured to perform the following steps:

[0179] Obtaining a user interest point pair, the user interest point pair including a user identifier and a content type identifier;

[0180] determining, by the user identifier, an overall content completion parameter of content of different content types, and a single-type completion parameter of target content corresponding to the content type identifier;

[0181] determining, based on the content type identifier, an overall user completion parameter of the target content by different users;

[0182] determining whether the user interest point is accurate according to the association relationship between the overall content completion parameter and the single-type completion parameter, and the association relationship between the overall user completion parameter and the single-type completion parameter.

[0183] Optionally, the CPU 1422 can also execute the method steps of any specific implementation manner of the data recognition method in the embodiments of the present application.

[0184] For the data recognition method described above, the embodiments of the present application further provide a terminal device for data recognition, so as to make the above data recognition method realize and apply in practice.

[0185] Referring to Figure 6 , Figure 6 a structural schematic diagram of a terminal device provided by the embodiments of the present application. For the convenience of description, only the parts related to the embodiments of the present application are shown, and the specific technical details not disclosed are referred to the method part of the embodiments of the present application. The terminal device can be any terminal device including a mobile phone, a tablet computer, a personal digital assistant (PDA), etc., and the terminal device is taken as a mobile phone as an example:

[0186] Figure 6 a block diagram of part of the structure of the mobile phone related to the terminal device provided by the embodiments of the present application. Referring to Figure 6 , the mobile phone includes a radio frequency (RF) circuit 1510, a memory 1520, an input unit 1530, a display unit 1540, a sensor 1550, an audio circuit 1560, a wireless fidelity (WiFi) module 1570, a processor 1580, and a power supply 1590, etc. Those skilled in the art can understand Figure 6 that the structure of the mobile phone shown in the figure does not constitute a limitation on the mobile phone, and can include more or less components than the figure, or combine certain components, or different component arrangements.

[0187] The specific introduction of each component of the mobile phone will be made below: Figure 6

[0188] ​The RF circuit 1510 can be used for receiving and sending signals in the process of information or communication, especially, receiving the downlink information from the base station and processing by the processor 1580. In addition, the uplink data is sent to the base station. Generally, the RF circuit 1510 includes, but is not limited to, an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier (LNA), a duplexer, etc. In addition, the RF circuit 1510 can also communicate with the network and other devices through wireless communication. The above-mentioned wireless communication can use any communication standard or protocol, including but not limited to global system for mobile communication (GSM), general packet radio service (GPRS), code division multiple access (CDMA), wideband code division multiple access (WCDMA), long term evolution (LTE), email, short messaging service (SMS), etc.

[0189] The memory 1520 can be used to store software programs and modules, and the processor 1580 can realize various functional applications and data processing of the mobile phone by running the software programs and modules stored in the memory 1520. The memory 1520 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory 1520 can include a high-speed random access memory, and can also include a non-volatile memory, for example, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state memory device.

[0190] The input unit 1530 can be used to receive inputted digital or character information, and to generate key signal input with respect to user setting of the mobile phone and control of function. Specifically, the input unit 1530 can include a touch panel 1531 and other input devices 1532. The touch panel 1531, also called a touch screen, can collect touch operations (such as operations of a user using a finger, a stylus, or any suitable object or accessory on or near the touch panel 1531) of the user on or near it, and drive corresponding connection devices according to a pre-set program. Optionally, the touch panel 1531 can include two parts, a touch detection device and a touch controller. The touch detection device detects the touch position of the user and detects signals caused by the touch operation, and transmits the signals to the touch controller; the touch controller receives the touch information from the touch detection device, converts it into touch coordinates, and sends it to the processor 1580, and can also receive commands from the processor 1580 and execute them. In addition, the touch panel 1531 can be implemented in various types such as resistive, capacitive, infrared, and surface acoustic wave. In addition to the touch panel 1531, the input unit 1530 can also include other input devices 1532. Specifically, the other input devices 1532 can include one or more of a physical keyboard, function keys (such as volume control keys, on / off keys, etc.), trackballs, mice, joysticks, etc.

[0191] The display unit 1540 can be used to display information input by the user or information provided to the user, as well as various menus of the mobile phone. The display unit 1540 can include a display panel 1541, which can be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), etc. Further, the touch panel 1531 can cover the display panel 1541, and when the touch panel 1531 detects a touch operation on or near it, it transmits to the processor 1580 to determine the type of touch event, and then the processor 1580 provides corresponding visual output on the display panel 1541 according to the type of touch event. Although in the above description, the touch panel 1531 and the display panel 1541 are realized as two independent components to realize the input and output functions of the mobile phone, in some embodiments, the touch panel 1531 and the display panel 1541 can be integrated to realize the input and output functions of the mobile phone. Figure 6

[0192] ​The phone can also include at least one sensor 1550, such as an optical sensor, a motion sensor, and other sensors. Specifically, the optical sensor can include an ambient light sensor to adjust the brightness of the display panel 1541 according to the brightness of ambient light, and a proximity sensor to turn off the display panel 1541 and / or the backlight when the phone is moved to the ear. As one of the motion sensors, the accelerometer sensor can detect the magnitude of acceleration in each direction (generally three axes), and when at rest, it can detect the magnitude and direction of gravity, which can be used for applications that identify the phone posture (such as landscape / portrait switching, related games, magnetometer posture calibration), vibration recognition related functions (such as pedometer, tapping), and the like. As for other sensors that the phone can also be configured, such as a gyroscope, a barometer, a hygrometer, a thermometer, an infrared sensor, and the like, they will not be described here.

[0193] The audio circuit 1560, the speaker 1561, and the microphone 1562 can provide an audio interface between the user and the phone. The audio circuit 1560 can convert the received audio data into an electrical signal, transmit it to the speaker 1561, and convert it into a sound signal output by the speaker 1561; on the other hand, the microphone 1562 converts the collected sound signal into an electrical signal, which is received by the audio circuit 1560 and converted into audio data, which is then processed by the processor 1580 and transmitted to, for example, another phone via the RF circuit 1510, or output to the memory 1520 for further processing.

[0194] WiFi is a short-range wireless transmission technology. The phone can help users send and receive emails, browse web pages, and access streaming media through the WiFi module 1570, which provides users with wireless broadband Internet access. Although Figure 6 Although the WiFi module 1570 is shown, it is understood that it does not belong to the necessary components of the phone and can be omitted as needed without changing the essence of the application.

[0195] The processor 1580 is the control center of the phone, which connects all parts of the phone through various interfaces and lines, and performs various functions and processes data of the phone by running or executing software programs and / or modules stored in the memory 1520 and calling data stored in the memory 1520. Optionally, the processor 1580 can include one or more processing units; preferably, the processor 1580 can integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communication. It is understood that the above-mentioned modem processor can also not be integrated into the processor 1580.

[0196] The mobile phone further includes a power supply 1590 (such as a battery) for supplying power to various components. Preferably, the power supply is logically connected to the processor 1580 through a power management system, so that the power management system can manage charging, discharging, and power consumption management.

[0197] Although not shown, the mobile phone can further include a camera, a Bluetooth module, and the like, which will not be described herein.

[0198] In the embodiments of the present application, the memory 1520 included in the mobile phone can store program code and transmit the program code to the processor.

[0199] The processor 1580 included in the mobile phone can execute the data recognition method provided in the above-described embodiments according to instructions in the program code.

[0200] The embodiments of the present application further provide a computer readable storage medium for storing a computer program, and the computer program is used to execute the data recognition method provided in the above-described embodiments.

[0201] The embodiments of the present application further provide a computer program product or a computer program, and the computer program product or the computer program includes computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device executes the data recognition method provided in various optional implementation manners of the above-described aspects.

[0202] Those skilled in the art can understand that all or part of the steps of the above-described method embodiments can be completed by program instruction related hardware, and the foregoing program can be stored in a computer readable storage medium. When the program is executed, the program executes steps including the above-described method embodiments. The foregoing storage medium can be at least one of the following media: a read-only memory (English: read-only memory, abbreviation: ROM), a RAM, a magnetic disk or an optical disk, and various media that can store program codes.

[0203] It should be noted that each of the implementations described in the specification is just one implementation and can not be implemented to the exclusion of other implementations. In other words, not all implementations described are necessarily mutually exclusive, and the specific order or hierarchy of steps in the implementations can be different from that described in the specification. Nor is the specification limited to these implementations or ordering of steps. In some implementations, steps can not be performed in the order noted in the implementation. For example, steps can be performed in an order different than presented in this document. Further, words such as "thereon," "thereby," "wherein," or "whereby" are not necessarily used in a causal relationship; for example, merely because step B is described after step A does not mean that step A is a cause of step B.

[0204] The above description is only specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A data identification method, characterized in that: The method comprises: Obtaining a user interest point pair, wherein the user interest point pair includes a user identifier and a content type identifier; Determining, by using the user identifier, an overall content completion parameter for content of different content types, and a single-type completion parameter for target content corresponding to the content type identifier, wherein the overall content completion parameter is used to identify an average playback completion of content of different content types by using the user identifier, and the single-type completion parameter is used to identify an average playback completion of content corresponding to the content type identifier by using the user identifier; Based on the content type identifier, determining an overall user completion parameter for the target content by different users, the overall user completion parameter being used to identify an average playback completion status of the content corresponding to the content type identifier by different users; If the overall content completion parameter is greater than the single type completion parameter, and the overall user completion parameter is greater than the single type completion parameter, it is determined that the user interest point pair is inaccurate; If the overall content completion parameter is less than or equal to the single-type completion parameter, and the overall user completion parameter is less than or equal to the single-type completion parameter, it is determined that the user interest point pair is accurate.

2. The method according to claim 1, characterized in that The obtaining of user interest point pairs includes: Distributing N contents corresponding to the content type identifier to M users, where N and M are integers greater than or equal to 1; The user interest point pairs are determined according to the content completion status of the M users on the distributed content.

3. The method according to claim 1, characterized in that After obtaining an accuracy result for identifying whether the user interest point pair is accurate, the method further includes: Obtain the user interest point pairs to be identified; The accuracy of the user interest point pairs to be identified is determined based on the distribution relationship between the user interest point pairs in the vector space and the user interest point pairs, where the vector space is constructed based on the vectorized representation of the user interest point pairs and the accuracy results of the user interest point pairs.

4. The method according to claim 3, characterized in that The method further comprises: According to the accuracy of the to-be-identified user interest point pair, the weight of the to-be-identified user interest point pair in the interest portrait of the user to which it belongs is adjusted.

5. The method according to claim 1, wherein The method further comprises: Determining a first average of preferences of different users for the target content, and a second average of preferences of the same user for content of different content types; determining a preference parameter according to the first preference mean and the second preference mean; Determining a first average exposure value of different users to the target content, and a second average exposure value of the same user to content of different content types; determining an exposure parameter according to the first exposure mean and the second exposure mean; The exposure strategy of the target content is regulated according to the exposure parameter and the preference parameter.

6. A data recognition device, characterized in that: The device includes: an acquisition unit, a determination unit and a judgment unit; The acquiring unit is configured to acquire a user interest point pair, wherein the user interest point pair includes a user identifier and a content type identifier; The determining unit is configured to determine, using the user identifier, an overall content completion parameter for content of different content types, and a single-type completion parameter for target content corresponding to the content type identifier, wherein the overall content completion parameter is used to identify an average playback completion of content of different content types using the user identifier, and the single-type completion parameter is used to identify an average playback completion of content corresponding to the content type identifier using the user identifier; Based on the content type identifier, determining an overall user completion parameter for the target content by different users, the overall user completion parameter being used to identify an average playback completion status of the content corresponding to the content type identifier by different users; The judgment unit is used to determine whether the user interest point pair is accurate based on the association relationship between the overall content completion parameter and the single type completion parameter, and the association relationship between the overall user completion parameter and the single type completion parameter. The judging unit is configured to: If the overall content completion parameter is greater than the single type completion parameter, and the overall user completion parameter is greater than the single type completion parameter, it is determined that the user interest point pair is inaccurate; If the overall content completion parameter is less than or equal to the single-type completion parameter, and the overall user completion parameter is less than or equal to the single-type completion parameter, it is determined that the user interest point pair is accurate.

7. The device according to claim 6, characterized in that The acquisition unit is configured to: Distributing N contents corresponding to the content type identifier to M users, where N and M are integers greater than or equal to 1; The user interest point pairs are determined according to the content completion status of the M users on the distributed content.

8. The device according to claim 6, characterized in that After obtaining the accuracy result for identifying whether the user interest point pair is accurate, the apparatus is further configured to: Determine the vectorized representation of the user's interest point pairs; constructing a vector space according to the vectorized representation and the accuracy results of the user interest point pairs; Obtain a pair of user interest points to be identified, and determine the accuracy of the pair of user interest points to be identified based on a distribution relationship between the pair of user interest points to be identified and the user interest point pairs in a vector space, where the vector space is constructed based on a vectorized representation of the user interest point pairs and an accuracy result of the user interest point pairs.

9. The device according to claim 8, characterized in that The device is also used for: According to the accuracy of the to-be-identified user interest point pair, the weight of the to-be-identified user interest point pair in the interest portrait of the user to which it belongs is adjusted.

10. The device according to claim 6, characterized in that The device is also used for: Determining a first average of preferences of different users for the target content, and a second average of preferences of the same user for content of different content types; determining a preference parameter according to the first preference mean and the second preference mean; Determining a first average exposure value of different users to the target content, and a second average exposure value of the same user to content of different content types; determining an exposure parameter according to the first exposure mean and the second exposure mean; The exposure strategy of the target content is regulated according to the exposure parameter and the preference parameter.

11. A computer device, characterized in that: The device includes a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is configured to execute the method according to any one of claims 1 to 5 according to instructions in the program code.

12. A computer-readable storage medium, characterized in that The computer-readable storage medium is used to store a computer program, and the computer program is used to execute the method according to any one of claims 1 to 5.

13. A computer program product, characterized in that The computer program product comprises instructions, which, when executed on a computer device, cause the computer device to perform the method according to any one of claims 1 to 5 .

Images