User portrait construction method, related device, equipment and storage medium

Abstract

The invention discloses a user portrait construction method, a related device, equipment and a storage medium, and the user portrait construction method comprises the steps: obtaining a plurality of historical interaction data clusters for a target user to interact with a target interaction system, and obtaining a plurality of historical interaction data clusters based on the interaction operation data collected when the target user interacts with the target interaction system at present; extracting a current interaction data cluster; wherein any interaction data cluster comprises attribute data of the interaction operation, and the attribute data of the interaction operation comprises an object name of an interaction object of the interaction operation in the target interaction system and an attribute value of at least one attribute type in object types; selecting the historical interaction data cluster as a target interaction data cluster based on the similarity of attribute data with each historical interaction data cluster in the current interaction data cluster; and constructing a user portrait of the target user based on the current interaction data cluster and the target interaction data cluster. According to the scheme, the accuracy of user portrait construction can be improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a user portrait construction method and related devices, equipment and storage media. Background Art

[0002] With the development of Internet technology, in order to realize the personalized functions of interactive systems that interact with users and enable the interactive systems to better serve users, user portrait construction technology has received more and more attention.

[0003] In the prior art, user profiles are usually constructed based on user interaction data collected within a preset time period. However, since users may have different interests and preferences when interacting with the interactive system, the interactive content recommended based on the constructed user profiles usually lags behind, which weakens the accuracy of user profile construction. In view of this, how to improve the accuracy of user profile construction has become an urgent problem to be solved. Summary of the invention

[0004] The main technical problem solved by this application is to provide a user portrait construction method and related devices, equipment and storage media, which can improve the accuracy of user portrait construction.

[0005] In order to solve the above technical problems, the first aspect of the present application provides a user portrait construction method, including: obtaining several historical interaction data clusters of the target user interacting with the target interaction system, and extracting the current interaction data cluster based on the interaction operation data collected when the target user currently interacts with the target interaction system; wherein any interaction data cluster includes attribute data of the interaction operation, and the attribute data of the interaction operation includes the object name of the interaction object in the target interaction system and the attribute value of at least one attribute type in the object category of the interaction operation; based on the similarity of the attribute data in the current interaction data cluster with each historical interaction data cluster, selecting the historical interaction data cluster as the target interaction data cluster; based on the current interaction data cluster and the target interaction data cluster, constructing a user portrait of the target user.

[0006] In order to solve the above technical problems, the second aspect of the present application provides a user portrait construction device, including: an acquisition module, a selection module and a construction module, the acquisition module is used to acquire several historical interaction data clusters of the target user interacting with the target interaction system, and extract the current interaction data cluster based on the interaction operation data collected when the target user currently interacts with the target interaction system; wherein any interaction data cluster includes attribute data of the interaction operation, and the attribute data of the interaction operation includes the object name of the interaction object in the target interaction system and the attribute value of at least one attribute type in the object category; the selection module is used to select the historical interaction data cluster as the target interaction data cluster based on the similarity of the attribute data in the current interaction data cluster with each historical interaction data cluster; the construction module is used to construct a user portrait of the target user based on the current interaction data cluster and the target interaction data cluster.

[0007] In order to solve the above-mentioned technical problems, the third aspect of the present application provides an electronic device, comprising a memory and a processor coupled to each other, wherein program instructions are stored in the memory, and the processor is used to execute the program instructions to implement the user portrait construction method in the above-mentioned first aspect.

[0008] In order to solve the above technical problems, the fourth aspect of the present application provides a computer-readable storage medium storing program instructions that can be executed by a processor, and the program instructions are used to implement the user portrait construction method described in the first aspect above.

[0009] The above scheme obtains several historical interaction data clusters of the target user interacting with the target interaction system, and extracts the current interaction data cluster based on the interaction operation data collected when the target user currently interacts with the target interaction system. Any interaction data cluster includes the attribute data of the interaction operation, and the attribute data of the interaction operation includes the object name of the interaction object in the target interaction system and the attribute value of at least one attribute type in the object category. Based on the similarity of the attribute data between the current interaction data cluster and each historical interaction data cluster, the historical interaction data cluster is selected as the target interaction data cluster. Based on the current interaction data cluster and the target interaction data cluster, the user profile of the target user is constructed. Therefore, the historical interaction data cluster associated with the current interaction data cluster when the target user interacts with the target interaction system is selected as the target interaction data cluster. The target interaction data cluster can reflect the behavior change trend and related interest preferences of the target user in the current interaction with the target interaction system. The target interaction data cluster and the current interaction data cluster are used together as reference information for constructing the user profile, which can provide more information about the interaction behavior of the target user when interacting with the target interaction system on the premise of depicting the short-term user profile of the target user. Therefore, the accuracy of user profile construction can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1This is a flow chart of an embodiment of a method for constructing a user portrait of the present application; Figure 2 This is a schematic diagram of the framework of an embodiment of a user portrait construction device of the present application; Figure 3 It is a schematic diagram of the framework of an embodiment of the electronic device of the present application; Figure 4 It is a schematic diagram of a framework of an embodiment of a computer-readable storage medium of the present application. DETAILED DESCRIPTION

[0011] The scheme of the embodiment of the present application is described in detail below in conjunction with the drawings of the specification.

[0012] In the following description, for the purpose of explanation rather than limitation, specific details such as specific system structures, interfaces, and technologies are provided to facilitate a thorough understanding of the present application.

[0013] If the technical solution of this application involves personal information, the product using the technical solution of this application has clearly informed the personal information processing rules and obtained the individual's voluntary consent before processing the personal information. If the technical solution of this application involves sensitive personal information, the product using the technical solution of this application has obtained the individual's separate consent before processing the sensitive personal information, and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, clear and prominent signs are set to inform that the personal information collection scope has been entered and personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that he or she agrees to the collection of his or her personal information; or on the device that processes personal information, the personal information processing rules are notified by obvious signs / information, and the individual's authorization is obtained through pop-up information or by asking the individual to upload his or her personal information; among them, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

[0014] The terms "system" and "network" are often used interchangeably in this article. The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, the fragment " / " in this article generally indicates that the associated objects before and after are in an "or" relationship. In addition, "many" in this article means two or more than two.

[0015] See also Figure 1 , Figure 1 This is a flow chart of an embodiment of the user portrait construction method of the present application. Specifically, it may include the following steps: Step S10: obtaining a plurality of historical interaction data clusters of the target user interacting with the target interaction system, and extracting a current interaction data cluster based on the interaction operation data collected when the target user currently interacts with the target interaction system.

[0016] In the embodiments of the present disclosure, the target interaction system is a hardware device or application software that can interact with the target user. For example, the hardware device may specifically include but is not limited to: smart phones, laptops, tablet computers, learning machines, etc., and the application software may specifically include but is not limited to video playback software, music playback software, social software, etc., which are not limited here.

[0017] In one implementation scenario, any interaction data cluster includes attribute data of the interaction operation, that is, both the historical interaction data cluster and the current interaction data cluster include attribute data of the interaction operation, and the attribute data of the interaction operation includes the object name of the interaction object in the target interaction system, and the attribute value of at least one attribute type in the object category. For example, the target user chooses to play a certain movie during the interaction with the target interaction system, and the attribute data of this interaction operation includes the object name of the movie being played, and the attribute value of at least one of the object categories of the movie being played, such as the attribute value of the object name is "Movie A", and the attribute value of the object category is "Comedy" or "Drama". For another example, the target user chooses to play a certain song during the interaction with the target interaction system, and the attribute data of this interaction operation includes at least the object name of the song being played, and the attribute value of at least one of the object categories of the song being played, such as the attribute value of the object name is "Song B", and the attribute value of the object category is "Quiet" or "Folk Song". It can be understood that the attribute value of the object category is at least one, and in different implementation scenarios, the number of object type attribute values ​​is not limited in this application.

[0018] In a specific implementation scenario, the interactive operation data in the interactive data cluster is the attribute value of the interactive operation attribute data, and the attribute value of the interactive operation attribute data is used to characterize the specific content of this interactive operation. The attribute data of the interactive operation also includes interactive behavior, interaction start time, interaction duration, etc. For example, the attribute value of the interactive behavior is "start playing", "stop playing", etc., the attribute value of the interaction start time is "xx / xx / xx / xx:xx:xx", etc., the attribute value of the interaction duration is "one hour", "one minute", etc. The number of attribute data of the interactive operation is related to the type of the interactive operation, and the specific number of attribute data of the interactive operation is not limited in this application.

[0019] In a specific implementation scenario, the interactive operation data is obtained by extracting fields from relevant logs of the interaction between the target user and the target interactive system. The logs store data on interactive events, and the stored data is streaming data. Streaming data is a set of sequential, large, fast, and continuously arriving data sequences, which can be regarded as a dynamic data set that grows infinitely over time. After obtaining several historical interaction data clusters of the target user interacting with the target interactive system, and extracting the current interaction data cluster based on the interactive operation data collected when the target user currently interacts with the target interactive system, the user name of the target user can be obtained, and the interactive operation data about the target user in the relevant logs of the interaction with the target system can be obtained by filtering the fields. In some implementation scenarios, such as the scenario where the target user uses a learning machine, the data of a targeted software used by the target user in the process of interacting with the learning machine can be obtained as the interactive operation data, that is, the interactive operation data about the targeted software in the relevant logs of the interaction with the target system can be obtained by filtering the fields, which is not limited in the present application.

[0020] In a specific implementation scenario, considering that the target user's interest in a series of interactive behavior sequences is focused, therefore, several historical interactive operation data of the target user interacting with the target interactive system are obtained in sequence, and the generation time interval between two consecutive sets of historical interactive operation data is determined in sequence. When the generation time interval is greater than a preset time threshold, the latter set of historical interactive operation data in the two consecutive sets of historical interactive operation data is added to a new historical interactive data cluster. When the generation time interval is not greater than the preset time threshold, the latter set of historical interactive operation data in the two consecutive sets of historical interactive operation data is added to the same historical interactive data cluster as the former set of historical interactive operation data in the historical interactive operation data, so as to obtain several historical interactive data clusters. It can be understood that the historical interactive operation data collected by the first sorting is divided into the initial historical interactive data cluster.

[0021] In a specific implementation scenario, after obtaining several historical interaction data clusters, the number of interactive operations contained in each historical interaction data cluster is detected, and the historical interaction data clusters with a number less than the minimum number threshold and the historical interaction data clusters with a number greater than the maximum number threshold are deleted, and the historical interaction data clusters that may have abnormalities are filtered out to further improve the accuracy of user portrait construction. It should be noted that the specific values ​​of the minimum number threshold and the maximum number threshold are not limited in this application.

[0022] In a specific implementation scenario, after obtaining the interaction operation data collected when the target user currently interacts with the target interaction system, it is determined based on a preset time interval whether to divide the currently collected interaction operation data and the last collected interaction operation data into the same current interaction data cluster; when the time interval between the currently collected interaction operation data and the last collected interaction operation data is not greater than the preset time interval, the currently collected interaction operation data is divided into the same current interaction data cluster as the last collected interaction operation data; when the time interval between the currently collected interaction operation data and the last collected interaction operation data is greater than the preset time interval, the currently collected interaction operation data is divided into a new current interaction data cluster.

[0023] In a specific implementation scenario, in order to avoid the situation where the current interaction data cluster contains too little interaction operation data to reflect the current interest level of the target user, when the interaction operation data in the latest generated current interaction data cluster does not meet the preset number, the step of extracting the current interaction data cluster based on the interaction operation data collected when the target user currently interacts with the target interaction system is returned to, until the number of interaction operation data in the current interaction data cluster is not less than the preset number, executing the step of selecting the historical interaction data cluster as the target interaction data cluster based on the similarity of attribute data between the current interaction data cluster and each historical interaction data cluster, and subsequent steps.

[0024] It should be noted that the specific value of the preset time interval is not limited in the present application. Specifically, the preset time interval is a fixed value set manually, or a dynamically changing value, and the value of the preset time interval is related to the target interaction system. For example, due to the different frequencies of use of the learning machine and the smart phone, the preset time interval set for the learning machine is greater than the time interval set for the smart phone.

[0025] In a specific implementation scenario, after constructing several historical interaction data clusters based on several historical interaction operation data of the target user interacting with the target interaction system, the historical interaction data clusters are stored. Specifically, the historical interaction data clusters can be constructed and stored based on a preset time period. For example, with one day as a cycle, the historical interaction operation data of the previous day is obtained on the same day and the historical interaction data clusters are constructed and stored.

[0026] Step S20: Based on the similarity of the attribute data between the current interaction data cluster and each historical interaction data cluster, a historical interaction data cluster is selected as a target interaction data cluster.

[0027] In one implementation scenario, based on the attribute data of each interactive operation in the current interactive data cluster, the first interactive feature is extracted, and based on the attribute data of each interactive operation in the historical interactive data cluster, the second interactive feature is extracted, and based on the similarity between the first interactive feature and each second interactive feature, the historical interactive data cluster is selected as the target interactive data cluster. In the above scheme, the historical interactive data cluster associated with the current interactive data cluster when the target user interacts with the target interactive system is selected as the target interactive data cluster. The target interactive data cluster can reflect the behavioral change trend and related interest preferences of the target user in the current interaction with the target interactive system, so it can further improve the accuracy of the subsequent user portrait construction.

[0028] In a specific implementation scenario, the attribute data of each interactive operation in the current interactive data cluster may be repeated. For example, the interactive operations generated by the target user interacting with the target interactive system in the current interactive data cluster include "target user plays movie A" and "target user plays movie B". The attribute values ​​representing the object categories of the interactive objects in "target user plays movie A" include "romance", "plot", and "comedy". The attribute values ​​representing the object categories of the interactive objects in "target user plays movie B" include "action" and "comedy". That is, the attribute values ​​representing the object categories of the interactive objects in the current interactive data cluster include "romance", "plot", "comedy", "action", and "comedy". It can be understood that the attribute data of each interactive operation in the historical interactive data cluster may also be repeated. For the sake of brevity, it will not be repeated here.

[0029] In a specific implementation scenario, based on the first attribute value of the attribute data of each interactive operation in the current interactive data cluster, a text vector about the first attribute value field is obtained as the first interactive feature, and based on the second attribute value of the attribute data of each interactive operation in the historical interactive data cluster, a text vector about the second attribute value field is obtained as the second interactive feature. Specifically, the first attribute value field includes "comedy", "comedy", and "plot", the second attribute value field in a certain historical interactive data cluster includes "suspense", "mainland", and "action", and the second attribute value field in another historical interactive data cluster includes "comedy", "love", "mainland", and "plot". A first text vector is generated based on "comedy", "plot", and "mainland" as the first interactive feature, a second text vector generated based on "suspense", "mainland", and "action" is used as a second interactive feature, and a third text vector is generated based on "comedy", "love", and "mainland" as another second interactive feature. The similarity between the first text vector and the second text vector and the similarity between the first text vector and the third text vector are respectively obtained, and a similarity comparison is performed, and the historical interactive data cluster corresponding to the larger similarity is selected as the target interactive data cluster.

[0030] In a specific implementation scenario, as a possible implementation method, a feature extraction model can be pre-trained, and the feature extraction model can include but is not limited to the network model of the Encoder-Decoder architecture, etc. In order to ensure the extraction accuracy of the feature extraction model as much as possible, sample interaction data clusters can be collected, and the sample interaction data clusters are annotated with real interaction features. On this basis, the sample interaction data clusters can be processed based on the feature extraction model to obtain the predicted interaction features of the sample interaction data clusters, so that the network parameters of the feature extraction model can be adjusted based on the difference between the real interaction features and the predicted interaction features until the feature extraction model training converges, and the current interaction data cluster and each historical interaction data cluster can be processed based on the converged feature extraction model to obtain the audio features of the target speech segment. It should be noted that the specific processing process of the feature extraction model can refer to the technical details such as the network model of the Encoder-Decoder architecture, which will not be repeated here.

[0031] In another specific implementation scenario, based on the attribute values ​​of the same attribute type involved in each interactive operation in the current interactive data cluster, the first sub-feature of the corresponding attribute type is extracted, and the first sub-feature includes weight factors of different attribute values ​​of the same attribute type involved in the current interactive data cluster. For example, the attribute values ​​of the object categories representing the interactive objects in the current interactive data cluster include "love", "plot", "comedy", "action", and "comedy", and the weight of "love" is 0.2, the weight of "plot" is 0.2, the weight of "comedy" is 0.4, and the weight of "action" is 0.2. Based on the attribute values ​​of the corresponding attribute type and the weights of the attribute values, the first sub-feature is generated, and the first sub-feature includes weight factors of different attribute values ​​of the same attribute type involved in the current interactive data cluster. Based on the attribute values ​​of the same attribute type involved in each interactive operation in the historical interactive data cluster, the second sub-feature of the corresponding attribute type is extracted, and the second sub-feature includes weight factors of different attribute values ​​of the same attribute type involved in the historical interactive data cluster. Based on the similarity between the first sub-feature and the second sub-feature of the same attribute type, the historical interactive data cluster is selected as the target interactive data cluster determined under the consideration dimension of the corresponding attribute type. The above scheme selects the historical interaction data cluster associated with the current interaction data cluster when the target user interacts with the target interaction system as the target interaction data cluster based on the similarity between the first sub-feature and the second sub-feature of the same attribute type. The target interaction data cluster can reflect the behavior change trend and related interest preferences of the target user in the current interaction with the target interaction system, and thus can further improve the accuracy of subsequent user portrait construction.

[0032] In a specific implementation scenario, based on the weight factor of the same attribute value in the first sub-feature and the second sub-feature of the same attribute type, the first local feature in the first sub-feature of the corresponding attribute type and the second local feature in the second sub-feature of the corresponding attribute type are extracted, and based on the similarity between the first local feature and the second local feature of the same attribute type, the historical interaction data cluster is selected as the target interaction data cluster determined under the consideration dimension of the corresponding attribute type.

[0033] In a specific implementation scenario, the weight factors of different attribute values ​​in the first sub-feature are positively correlated with the occurrence frequency of each attribute value in the current interaction data cluster, that is, the higher the occurrence frequency of the attribute value in the current interaction data cluster, the larger the corresponding weight factor, and / or, the weight factors of different attribute values ​​in the second sub-feature are positively correlated with the occurrence frequency of each attribute value in the historical interaction data cluster, that is, the higher the occurrence frequency of the attribute value in the historical interaction data cluster, the larger the corresponding weight factor.

[0034] In a specific implementation scenario, before extracting the second sub-feature of the corresponding attribute type based on the attribute values ​​of the same attribute type involved in each interactive operation in the historical interaction data cluster, the comparison result between the attribute values ​​of the same attribute type in the historical interaction data cluster and the current interaction data cluster is detected, and the comparison result includes that the attribute values ​​of the same attribute type are at least partially the same, and the historical interaction data cluster is selected as the historical interaction data cluster to be extracted under the consideration dimension of the corresponding attribute type. The above scheme performs a preliminary screening based on the attribute values ​​of the attribute type under the consideration dimension of the corresponding attribute type before selecting the target interaction data cluster, thereby reducing the amount of data processing.

[0035] In a specific implementation scenario, based on the similarity of attribute data between the current interaction data cluster and each historical interaction data cluster, at least one historical interaction data cluster is selected as the target interaction data cluster. Specifically, the historical interaction data cluster corresponding to the highest similarity is selected as the target interaction data cluster, or a similarity threshold is set, and the historical interaction data cluster corresponding to a similarity not less than the similarity threshold is selected as the target interaction data cluster. It should be noted that the above embodiment is only a possible implementation method, and the number of target interaction data clusters is not limited in this application.

[0036] In a specific implementation scenario, the current interaction data cluster can build a retrieval index based on the attribute value of the object name of the interaction object in the target interaction system and the attribute value of the object label of the interaction object in the target interaction system. Similarly, the historical interaction data cluster can build a retrieval index based on the attribute value of the object name of the interaction object in the target interaction system and the attribute value of the object label of the interaction object in the target interaction system. Preliminary screening and similarity calculation are performed based on the retrieval index, thereby reducing the computing pressure of the data and improving the efficiency of user portrait construction.

[0037] Step S30: constructing a user profile of the target user based on the current interaction data cluster and the target interaction data cluster.

[0038] In the disclosed embodiments, the user portrait of the target user is used to analyze the data of the target user's social attributes, living habits, consumption behavior and other main information, and abstract the overall business picture of the target user through these data. The user portrait constructed in this application mainly includes the user's implicit characteristics, such as purpose of use, preference, needs and usage scenarios. It should be noted that the application collects the interaction data of the target user interacting with the target interactive system, constructs the user portrait, and performs personalized recommendations based on the user portrait and other related data operations are all performed after obtaining the target user data collection authorization.

[0039] In an implementation scenario, the historical interaction data cluster associated with the current interaction data cluster when the target user interacts with the target interaction system is selected as the target interaction data cluster. The target interaction data cluster can reflect the behavior change trend and related interest preferences of the target user in the current interaction with the target interaction system. The target interaction data cluster and the current interaction data cluster are used together as reference information for constructing a user portrait. On the premise of depicting the short-term user portrait of the target user, more information about the interaction behavior of the target user when interacting with the target interaction system can be provided. Therefore, the accuracy of user portrait construction can be improved.

[0040] In one implementation scenario, when the attribute data of the interactive operation includes the object name and the attribute value of the object category of the interactive object of the interactive operation in the target interactive system, a first sub-portrait is generated based on the current interactive data cluster and the historical interactive data cluster as the target interactive data cluster determined under the consideration dimension of the corresponding object name, and a second sub-portrait is generated based on the current interactive data cluster and the historical interactive data cluster as the target interactive data cluster determined under the consideration dimension of the corresponding object category, and a user portrait of the target user is constructed based on the first sub-portrait and the second sub-portrait. The above scheme can improve the data richness used for user portrait construction and improve the accuracy of user portrait construction under the premise of considering the target user's interaction habits as much as possible.

[0041] In one implementation scenario, after selecting a historical interaction data cluster as a target interaction data cluster based on the similarity of attribute data between the current interaction data cluster and each historical interaction data cluster, the latest selection time when each historical interaction data cluster was selected as the target interaction data cluster is obtained. When the latest selection time of the historical interaction data cluster does not meet the preset time condition, for example, the latest selection time has exceeded one year, the historical interaction data cluster is deleted. The above scheme promptly deletes historical interaction data clusters that have not been selected for a long time, ensures the regular cleaning of data storage, does not cause problems such as a decrease in matching selection efficiency and data crashes due to excessive data volume, and improves the effect of user portrait construction.

[0042] In one implementation scenario, after constructing a user portrait of the target user based on the current interaction data cluster and the target interaction data cluster, it is detected that new interaction operation data is generated when the target user is currently interacting with the target interaction system, and the step of extracting the current interaction data cluster based on the interaction operation data collected when the target user is currently interacting with the target interaction system is returned to update the user portrait of the target object. The above scheme selects the historical interaction data cluster associated with the current interaction data cluster when the target user interacts with the target interaction system as the target interaction data cluster. The target interaction data cluster can reflect the behavior change trend and related interest preferences of the target user's current interaction with the target interaction system. The target interaction data cluster and the current interaction data cluster are used together as reference information for constructing the user portrait. On the premise of portraying the short-term user portrait of the target user, it can provide more information about the interaction behavior of the target user when interacting with the target interaction system, and update the user portrait of the target user in real time. Therefore, the accuracy of user portrait construction can be improved.

[0043] In one implementation scenario, after constructing a user portrait of the target user using the target interactive system, relevant interactive content can be recommended to the target user based on the user portrait, such as recommended movies, songs, videos, etc., to satisfy the target user's interests as much as possible, consider the migration of the target user's interest preferences, and enhance the user's sense of use of the target interactive system.

[0044] The above scheme obtains several historical interaction data clusters of the target user interacting with the target interaction system, and extracts the current interaction data cluster based on the interaction operation data collected when the target user currently interacts with the target interaction system. Any interaction data cluster includes the attribute data of the interaction operation, and the attribute data of the interaction operation includes the object name of the interaction object in the target interaction system and the attribute value of at least one attribute type in the object category. Based on the similarity of the attribute data between the current interaction data cluster and each historical interaction data cluster, the historical interaction data cluster is selected as the target interaction data cluster. Based on the current interaction data cluster and the target interaction data cluster, the user profile of the target user is constructed. Therefore, the historical interaction data cluster associated with the current interaction data cluster when the target user interacts with the target interaction system is selected as the target interaction data cluster. The target interaction data cluster can reflect the behavior change trend and related interest preferences of the target user in the current interaction with the target interaction system. The target interaction data cluster and the current interaction data cluster are used together as reference information for constructing the user profile, which can provide more information about the interaction behavior of the target user when interacting with the target interaction system on the premise of depicting the short-term user profile of the target user. Therefore, the accuracy of user profile construction can be improved.

[0045] See also Figure 2 , Figure 2 2 is a schematic diagram of a framework of an embodiment of a user portrait construction device 20 of the present application. Figure 2 As shown, the user portrait construction device 20 includes: an acquisition module 21, a selection module 22 and a construction module 23, the acquisition module 21 is used to obtain several historical interaction data clusters of the target user interacting with the target interaction system, and extract the current interaction data cluster based on the interaction operation data collected when the target user currently interacts with the target interaction system; wherein any interaction data cluster includes attribute data of the interaction operation, and the attribute data of the interaction operation includes the object name of the interaction object in the target interaction system and the attribute value of at least one attribute type in the object category; the selection module 22 is used to select the historical interaction data cluster as the target interaction data cluster based on the similarity of the attribute data in the current interaction data cluster with each historical interaction data cluster; the construction module 23 is used to construct a user portrait of the target user based on the current interaction data cluster and the target interaction data cluster.

[0046] In the above scheme, the user portrait construction device 20 obtains several historical interaction data clusters of the target user interacting with the target interaction system, and extracts the current interaction data cluster based on the interaction operation data collected when the target user currently interacts with the target interaction system. Any interaction data cluster includes the attribute data of the interaction operation, and the attribute data of the interaction operation includes the object name of the interaction object in the target interaction system and the attribute value of at least one attribute type in the object category. Based on the similarity of the attribute data between the current interaction data cluster and each historical interaction data cluster, the historical interaction data cluster is selected as the target interaction data cluster. Based on the current interaction data cluster and the target interaction data cluster, the user portrait of the target user is constructed. Therefore, the historical interaction data cluster associated with the current interaction data cluster when the target user interacts with the target interaction system is selected as the target interaction data cluster. The target interaction data cluster can reflect the behavior change trend and related interest preference of the target user in the current interaction with the target interaction system. The target interaction data cluster and the current interaction data cluster are used together as reference information for constructing the user portrait, which can provide more information about the interaction behavior of the target user when interacting with the target interaction system on the premise of depicting the short-term user portrait of the target user. Therefore, the accuracy of user portrait construction can be improved.

[0047] In some disclosed embodiments, the selection module 22 also includes an interaction feature extraction module (not shown), which is used to extract a first interaction feature based on the attribute data of each interaction operation in the current interaction data cluster, and to extract a second interaction feature based on the attribute data of each interaction operation in the historical interaction data cluster; the selection module 22 also includes a feature similarity module (not shown), which is used to select a historical interaction data cluster as a target interaction data cluster based on the similarity between the first interaction feature and each second interaction feature.

[0048] In some disclosed embodiments, the interaction feature extraction module (not shown) also includes a first feature extraction module (not shown) for extracting a first sub-feature of the corresponding attribute type based on the attribute values ​​of the same attribute type involved in each interaction operation in the current interaction data cluster; wherein the first sub-feature includes a weight factor for different attribute values ​​of the same attribute type involved in the current interaction data cluster; the interaction feature extraction module (not shown) also includes a second feature extraction module (not shown) for extracting a second sub-feature of the corresponding attribute type based on the attribute values ​​of the same attribute type involved in each interaction operation in the historical interaction data cluster; wherein the second sub-feature includes a weight factor for different attribute values ​​of the same attribute type involved in the historical interaction data cluster; the feature similarity module (not shown) also includes a selection submodule (not shown) for selecting the historical interaction data cluster as the target interaction data cluster determined under the consideration dimension of the corresponding attribute type based on the similarity between the first sub-feature and the second sub-feature of the same attribute type.

[0049] In some disclosed embodiments, the selection submodule (not shown) further includes a local feature extraction module (not shown) for extracting a first local feature in the first subfeature of the corresponding attribute type and a second local feature in the second subfeature of the corresponding attribute type based on a weight factor of the same attribute value in the first subfeature and the second subfeature of the same attribute type; the selection submodule (not shown) further includes a local feature selection module (not shown) for selecting a historical interaction data cluster as a target interaction data cluster determined under the consideration dimension of the corresponding attribute type based on the similarity between the first local feature and the second local feature of the same attribute type.

[0050] In some disclosed embodiments, before extracting the second sub-feature of the corresponding attribute type based on the attribute values ​​of the same attribute type involved in each interactive operation in the historical interaction data cluster, the interaction feature extraction module (not shown) also includes a comparison detection module (not shown) for detecting the comparison results between the attribute values ​​of the same attribute type in the historical interaction data cluster and the current interaction data cluster; the interaction feature extraction module (not shown) also includes a screening module (not shown) for selecting the historical interaction data cluster as the historical interaction data cluster for extracting features under the consideration dimension of the corresponding attribute type in response to the comparison results including that the attribute values ​​of the same attribute type are at least partially identical.

[0051] In some disclosed embodiments, the acquisition module 21 also includes a historical data acquisition module (not shown), which is used to sequentially acquire a number of historical interaction operation data of the target user interacting with the target interaction system; the acquisition module 21 also includes a time interval judgment module (not shown), which is used to sequentially judge the generation time interval between two consecutive sets of historical interaction operation data; the acquisition module 21 also includes a data cluster construction module (not shown), which is used to add the latter set of historical interaction operation data in two consecutive sets of historical interaction operation data to a new historical interaction data cluster in response to the generation time interval being greater than a preset time threshold, so as to obtain a number of historical interaction data clusters.

[0052] In some disclosed embodiments, the construction module 23 also includes a sub-portrait generation module (not shown) for generating a first sub-portrait based on the current interaction data cluster and the historical interaction data cluster as the target interaction data cluster determined under the consideration dimension of the corresponding object name, and generating a second sub-portrait based on the current interaction data cluster and the historical interaction data cluster as the target interaction data cluster determined under the consideration dimension of the corresponding object category; the construction module 23 also includes a user portrait generation module (not shown) for constructing a user portrait of the target user based on the first sub-portrait and the second sub-portrait.

[0053] In some disclosed embodiments, the user portrait construction device 20 also includes a selection time acquisition module (not shown), which is used to obtain each historical interaction data cluster of the target user interacting with the target interaction system as the latest selection time of the target interaction data cluster; the user portrait construction device 20 also includes a data cluster deletion module (not shown), which is used to delete the historical interaction data cluster matching the latest selection time in response to the latest selection time not satisfying the preset time condition.

[0054] In some disclosed embodiments, after constructing a user portrait of the target user based on the current interaction data cluster and the target interaction data cluster, the user portrait construction device 20 also includes a portrait update module (not shown) for returning to execute the step of extracting the current interaction data cluster based on the interaction operation data collected when the target user is currently interacting with the target interaction system in response to detecting that new interaction operation data is generated when the target user is currently interacting with the target interaction system, so as to update the user portrait of the target object.

[0055] See also Figure 3 , Figure 3 : It is a schematic diagram of the framework of an embodiment of an electronic device 30 of the present application. The electronic device 30 includes a memory 31 and a processor 32. The memory 31 stores program instructions, and the processor 32 is used to execute the program instructions to implement the steps in any of the above-mentioned user portrait construction method embodiments. For details, please refer to the aforementioned disclosed embodiments, which will not be repeated here. The electronic device 30 may specifically include but is not limited to: a server, a smart phone, a laptop computer, a tablet computer, a self-service machine, etc., which are not limited here.

[0056] Specifically, the processor 32 is used to control itself and the memory 31 to implement the steps in any of the above-mentioned user portrait construction method embodiments. The processor 32 can also be called a CPU (Central Processing Unit). The processor 32 may be an integrated circuit chip with signal processing capabilities. The processor 32 can also be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field-programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. In addition, the processor 32 can be implemented by an integrated circuit chip.

[0057] In the above scheme, the electronic device 30 obtains several historical interaction data clusters of the target user interacting with the target interaction system, and extracts the current interaction data cluster based on the interaction operation data collected when the target user currently interacts with the target interaction system. Any interaction data cluster includes attribute data of the interaction operation, and the attribute data of the interaction operation includes the object name of the interactive object in the target interaction system and the attribute value of at least one attribute type in the object category. Based on the similarity of the attribute data between the current interaction data cluster and each historical interaction data cluster, the historical interaction data cluster is selected as the target interaction data cluster. Based on the current interaction data cluster and the target interaction data cluster, the user profile of the target user is constructed. Therefore, the historical interaction data cluster associated with the current interaction data cluster when the target user interacts with the target interaction system is selected as the target interaction data cluster. The target interaction data cluster can reflect the behavior change trend and related interest preference of the target user in the current interaction with the target interaction system. The target interaction data cluster and the current interaction data cluster are used together as reference information for constructing the user profile, which can provide more information about the interaction behavior of the target user when interacting with the target interaction system on the premise of depicting the short-term user profile of the target user. Therefore, the accuracy of user profile construction can be improved.

[0058] See also Figure 4 , Figure 4 4 is a schematic diagram of a framework of an embodiment of a computer-readable storage medium 40 of the present application. The computer-readable storage medium 40 stores program instructions 41 that can be executed by a processor, and the program instructions 41 are used to implement the steps in any of the above-mentioned user portrait construction method embodiments.

[0059] In the above scheme, the computer-readable storage medium 40 obtains several historical interaction data clusters of the target user interacting with the target interaction system, and extracts the current interaction data cluster based on the interaction operation data collected when the target user currently interacts with the target interaction system. Any interaction data cluster includes attribute data of the interaction operation, and the attribute data of the interaction operation includes the object name of the interaction object in the target interaction system and the attribute value of at least one attribute type in the object category. Based on the similarity of the attribute data between the current interaction data cluster and each historical interaction data cluster, the historical interaction data cluster is selected as the target interaction data cluster. Based on the current interaction data cluster and the target interaction data cluster, the user profile of the target user is constructed. Therefore, the historical interaction data cluster associated with the current interaction data cluster when the target user interacts with the target interaction system is selected as the target interaction data cluster. The target interaction data cluster can reflect the behavior change trend and related interest preference of the target user in the current interaction with the target interaction system. The target interaction data cluster and the current interaction data cluster are used together as reference information for constructing the user profile, which can provide more information about the interaction behavior of the target user when interacting with the target interaction system on the premise of depicting the short-term user profile of the target user. Therefore, the accuracy of user profile construction can be improved.

[0060] In some embodiments, the functions or modules included in the device provided by the embodiments of the present disclosure can be used to execute the method described in the above method embodiments. The specific implementation can refer to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0061] The above description of various embodiments tends to emphasize the differences between the various embodiments. The same or similar aspects can be referenced to each other, and for the sake of brevity, they will not be repeated herein.

[0062] In the several embodiments provided in the present application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device implementation described above is only schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0063] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0064] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0065] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of each implementation method of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

Claims

1. A method for constructing a user portrait, characterized in that: include: Acquire several historical interaction data clusters of the target user interacting with the target interaction system, and extract the current interaction data cluster based on the interaction operation data collected when the target user currently interacts with the target interaction system; wherein any interaction data cluster includes attribute data of the interaction operation, and the attribute data of the interaction operation includes the object name of the interaction object in the target interaction system and the attribute value of at least one attribute type in the object category of the interaction operation; Based on the similarity of the attribute data between the current interaction data cluster and each of the historical interaction data clusters, selecting the historical interaction data cluster as a target interaction data cluster; A user profile of the target user is constructed based on the current interaction data cluster and the target interaction data cluster.

2. The method according to claim 1, characterized in that The selecting the historical interaction data cluster as the target interaction data cluster based on the similarity between the current interaction data cluster and each of the historical interaction data clusters regarding the attribute data comprises: Extracting a first interaction feature based on the attribute data of each interaction operation in the current interaction data cluster, and extracting a second interaction feature based on the attribute data of each interaction operation in the historical interaction data cluster; Based on the similarities between the first interaction feature and each of the second interaction features, the historical interaction data cluster is selected as a target interaction data cluster.

3. The method according to claim 2, characterized in that The extracting a first interaction feature based on the attribute data of each interaction operation in the current interaction data cluster includes: Extracting a first sub-feature corresponding to the attribute type based on the attribute values ​​of the same attribute type involved in each of the interactive operations in the current interactive data cluster; wherein the first sub-feature includes weight factors of different attribute values ​​of the same attribute type involved in the current interactive data cluster; The extracting a second interaction feature based on the attribute data of each interaction operation in the historical interaction data cluster includes: Extracting a second sub-feature corresponding to the attribute type based on the attribute values ​​of the same attribute type involved in each of the interaction operations in the historical interaction data cluster; wherein the second sub-feature includes weight factors of different attribute values ​​of the same attribute type involved in the historical interaction data cluster; The selecting the historical interaction data cluster as the target interaction data cluster based on the similarities between the first interaction feature and each of the second interaction features includes: Based on the similarity between the first sub-feature and the second sub-feature of the same attribute type, the historical interaction data cluster is selected as the target interaction data cluster determined under the consideration dimension corresponding to the attribute type.

4. The method according to claim 3, characterized in that The selecting the historical interaction data cluster as the target interaction data cluster determined under the consideration dimension corresponding to the attribute type based on the similarity between the first sub-feature and the second sub-feature of the same attribute type includes: Based on the weight factor of the same attribute value in the first sub-feature and the second sub-feature of the same attribute type, extracting a first local feature in the first sub-feature corresponding to the attribute type and a second local feature in the second sub-feature corresponding to the attribute type; Based on the similarity between the first local feature and the second local feature of the same attribute type, the historical interaction data cluster is selected as the target interaction data cluster determined under the consideration dimension corresponding to the attribute type.

5. The method according to claim 3, characterized in that: The weight factors of different attribute values ​​in the first sub-feature are positively correlated with the occurrence frequencies of each attribute value in the current interaction data cluster; And / or, the weight factors of different attribute values ​​in the second sub-feature are positively correlated with the occurrence frequency of each attribute value in the historical interaction data cluster.

6. The method according to claim 3, characterized in that Before extracting the second sub-feature corresponding to the attribute type based on the attribute values ​​of the same attribute type involved in each of the interaction operations in the historical interaction data cluster, the method further includes: Detecting a comparison result between attribute values ​​of the same attribute type in the historical interaction data cluster and the current interaction data cluster; In response to the comparison result including that attribute values ​​of the same attribute type are at least partially identical, the historical interaction data cluster is selected as the historical interaction data cluster to be extracted features under the consideration dimension corresponding to the attribute type.

7. The method according to claim 1, characterized in that The step of obtaining several historical interaction data clusters of the target user interacting with the target interaction system includes: Sequentially acquiring a number of historical interaction operation data of the target user interacting with the target interaction system; sequentially determining the generation time interval between two consecutive groups of the historical interactive operation data; In response to the generation time interval being greater than a preset time threshold, the latter group of the historical interaction operation data in the two consecutive groups of the historical interaction operation data is added to a new historical interaction data cluster to obtain a plurality of the historical interaction data clusters.

8. The method according to claim 1, characterized in that: The constructing a user profile of the target user based on the current interaction data cluster and the target interaction data cluster includes: Generate a first sub-portrait based on the current interaction data cluster and the historical interaction data cluster as target interaction data clusters determined under the consideration dimension corresponding to the object name, and generate a second sub-portrait based on the current interaction data cluster and the historical interaction data cluster as target interaction data clusters determined under the consideration dimension corresponding to the object category; Based on the first sub-portrait and the second sub-portrait, the user portrait of the target user is constructed.

9. The method according to claim 1, characterized in that: The method further comprises: Acquire the latest selection time when each of the historical interaction data clusters is selected as the target interaction data cluster; In response to the latest selection time of the historical interaction data cluster not satisfying a preset time condition, deleting the historical interaction data cluster.

10. The method according to any one of claims 1 to 9, characterized in that: After constructing the user portrait of the target user based on the current interaction data cluster and the target interaction data cluster, the method further includes: In response to detecting that new interaction operation data is generated when the target user is currently interacting with the target interaction system, return to execute the step of extracting the current interaction data cluster based on the interaction operation data collected when the target user is currently interacting with the target interaction system to update the user portrait of the target object.

11. A user portrait construction device, characterized in that: include: an acquisition module, used to acquire several historical interaction data clusters of the target user interacting with the target interaction system, and extract the current interaction data cluster based on the interaction operation data collected when the target user currently interacts with the target interaction system; wherein any interaction data cluster includes attribute data of the interaction operation, and the attribute data of the interaction operation includes the object name of the interaction object of the interaction operation in the target interaction system and the attribute value of at least one attribute type in the object category; A selection module, configured to select the historical interaction data cluster as a target interaction data cluster based on the similarity between the current interaction data cluster and each of the historical interaction data clusters regarding the attribute data; A construction module is used to construct a user portrait of the target user based on the current interaction data cluster and the target interaction data cluster.

12. An electronic device, characterized in that: It includes a memory and a processor coupled to each other, wherein the memory stores program instructions, and the processor is used to execute the program instructions to implement the user portrait method described in any one of claims 1 to 10.

13. A computer-readable storage medium having program instructions stored thereon, characterized in that: When the program instructions are executed by a processor, the user portrait construction method described in any one of claims 1 to 10 is implemented.

Images