An image data construction method and device, electronic equipment and storage medium

Abstract

The present disclosure provides a portrait data construction method and device, electronic equipment and storage medium, relates to the technical field of artificial intelligence, and in particular to the fields of natural language processing and data processing. The specific implementation scheme is: performing feature data processing on the to-be-processed data obtained from multiple data sources to obtain multi-dimensional feature data corresponding to the to-be-processed data, wherein the multi-dimensional feature data is used to represent multiple data types corresponding to the to-be-processed data; and constructing portrait data according to the multi-dimensional feature data. According to the present disclosure, portrait data for describing multi-dimensional features of data can be accurately constructed.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and in particular to the fields of natural language processing and data processing. Background Technology

[0002] User profiles are cognitive representations of users obtained through the analysis of data to be processed, and serve as the starting point for subsequent data processing such as data analysis, data modeling, and matching with business needs. These user profiles can be obtained through data statistics (by collecting and analyzing data such as users' social attributes, lifestyle habits, consumption behavior, and interests, etc.).

[0003] However, current methods for building user profiles are too broad, resulting in inaccurate profiles that cannot provide precise data support for subsequent data analysis, data modeling, and matching business needs. Summary of the Invention

[0004] This disclosure provides a method, apparatus, electronic device, and storage medium for constructing portrait data.

[0005] According to one aspect of this disclosure, a method for constructing profile data is provided, comprising:

[0006] Data processing is performed on the data to be processed obtained from multiple data sources to obtain multi-dimensional feature data corresponding to the data to be processed. The multi-dimensional feature data is used to characterize the various data types corresponding to the data to be processed.

[0007] Based on the multi-dimensional feature data, profile data is constructed.

[0008] According to another aspect of this disclosure, an apparatus for constructing portrait data is provided, comprising:

[0009] The acquisition unit is used to perform feature data processing on the data to be processed acquired from multiple data sources to obtain multi-dimensional feature data corresponding to the data to be processed. The multi-dimensional feature data is used to characterize the various data types corresponding to the data to be processed.

[0010] The portrait construction unit is used to construct portrait data based on the multi-dimensional feature data.

[0011] According to another aspect of this disclosure, an electronic device is provided, comprising:

[0012] At least one processor; and

[0013] The memory is communicatively connected to the at least one processor; wherein,

[0014] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method provided in any embodiment of this disclosure.

[0015] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions for causing the computer to perform the methods provided in any embodiment of this disclosure.

[0016] According to another aspect of this disclosure, a computer program product is provided, including computer instructions that, when executed by a processor, implement the methods provided in any embodiment of this disclosure.

[0017] By employing this disclosure, data to be processed can be obtained from multiple data sources, and feature data processing can be performed on the data to be processed to obtain multi-dimensional feature data corresponding to the data to be processed. The multi-dimensional feature data is used to characterize the various data types corresponding to the data to be processed, and profile data can be constructed based on the multi-dimensional feature data. Thus, profile data describing the multi-dimensional features of the data can be accurately constructed.

[0018] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0019] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0020] Figure 1 This is a schematic diagram of a distributed cluster processing scenario according to an embodiment of the present disclosure;

[0021] Figure 2 This is a flowchart illustrating a method for constructing portrait data according to an embodiment of this disclosure;

[0022] Figure 3 This is a schematic diagram of the image infrastructure in an application example according to an embodiment of this disclosure;

[0023] Figure 4 This is a schematic diagram of image updating and modeling in an application example according to an embodiment of this disclosure;

[0024] Figure 5 This is a schematic diagram of the composition structure of the image data construction apparatus according to an embodiment of the present disclosure;

[0025] Figure 6 This is a block diagram of an electronic device used to implement the method for constructing portrait data according to embodiments of the present disclosure. Detailed Implementation

[0026] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0027] In this document, the term "and / or" merely describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. The term "at least one" in this document indicates any combination of at least two of a plurality of elements. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C. The terms "first" and "second" in this document refer to and distinguish between multiple similar technical terms, not to restrict the order or to limit there to only two. For example, "first feature" and "second feature" refer to two categories / two features; the first feature can be one or more, and the second feature can also be one or more.

[0028] Furthermore, to better illustrate this disclosure, numerous specific details are set forth in the following detailed description. Those skilled in the art will understand that this disclosure can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art have not been described in detail in order to highlight the main points of this disclosure.

[0029] Taking document download as an example, a business scenario involving data processing, the user's primary need is to find and download documents of interest. User needs for documents are multifaceted, including content quality, timeliness, popularity, and professionalism. Currently, user profiles for document download scenarios only include basic features (such as the account used to log into the document database and previously downloaded documents), lacking more advanced features (such as ratings of the content quality of downloaded documents). Using advanced features would allow for a more accurate portrayal of the user profile.

[0030] To better meet users' diverse needs for document downloads, it is necessary to combine basic and advanced features to construct multi-dimensional profile data, thereby better serving users.

[0031] Considering that document download scenarios involve diverse user needs (e.g., liking or disliking one or more documents, preferred document types, and which age group prefers which document type), as well as diverse user requirements regarding the document itself (e.g., document content quality, timeliness, popularity, and professionalism in terms of literature or technology), basic features alone cannot adequately address all these diverse needs. Therefore, this disclosure combines basic and advanced features to construct multi-dimensional profile data (user profiles and document profiles) based on these multi-dimensional features. This improves the accuracy of the constructed profile data and provides precise data support for subsequent data analysis, data modeling, and matching business needs.

[0032] According to embodiments of this disclosure, Figure 1 This is a schematic diagram of a distributed cluster processing scenario according to an embodiment of the present disclosure. The distributed cluster system is an example of a cluster system, exemplarily describing how this distributed cluster system can be used to construct profile data. This disclosure is not limited to constructing profile data on a single machine or multiple machines; using distributed processing can further improve the accuracy of the constructed profile data. By acquiring profile data from multiple data sources and analyzing multi-dimensional feature data, multi-dimensional profile data is ultimately obtained. Figure 1 As shown, this distributed cluster system includes multiple nodes (such as server cluster 101, server 102, server cluster 103, server 104, and server 105; server 105 can also connect to electronic devices, such as mobile phone 1051 and desktop computer 1052). Multiple nodes, as well as multiple nodes and connected electronic devices, can jointly execute one or more profile data construction tasks. Optionally, the multiple nodes in this distributed cluster system can adopt a data-parallel profile data construction method, in which case multiple nodes can execute profile data construction tasks based on the same data processing method (such as executing the overall data processing flow, rather than a partial data processing flow). Alternatively, the multiple nodes in this distributed cluster system can also use different data processing methods (executing partial data processing flows in parallel on different nodes, and then integrating the results of multiple partial data processing flows) to execute profile data construction tasks. After each round of profile data construction tasks is completed, multiple nodes can exchange data (such as data synchronization) to update the data.

[0033] According to embodiments of this disclosure, a method for constructing portrait data is provided. Figure 2This is a flowchart illustrating a method for constructing portrait data according to an embodiment of this disclosure. This method can be applied to a portrait data construction apparatus. For example, the apparatus can be deployed on a terminal, server, or other processing device in a single-machine, multi-machine, or cluster system to perform portrait data construction and other processing. The terminal can be a user equipment (UE), mobile device, personal digital assistant (PDA), handheld device, computing device, in-vehicle device, wearable device, etc. In some possible implementations, the method can also be implemented by a processor calling computer-readable instructions stored in memory. Figure 2 As shown, this method is applied to Figure 1 In any node or electronic device (such as a mobile phone or desktop computer) of the cluster system shown, the following are included:

[0034] S201. Perform feature data processing on the data to be processed obtained from multiple data sources to obtain multi-dimensional feature data corresponding to the data to be processed. The multi-dimensional feature data is used to characterize the various data types corresponding to the data to be processed.

[0035] S202. Based on the multi-dimensional feature data, construct the profile data.

[0036] In one example of S201-S202, data to be processed can be obtained from multiple data sources (such as data sources storing document data, user data, registered / paid membership data, user behavior data, and login account authentication data). Data processing, such as extraction-transformation-loading (ETL), is then performed on this data to obtain multi-dimensional feature data corresponding to it (e.g., basic features including "consumption features" and advanced features including "production features and / or interaction features"). Profile data can be constructed based on this multi-dimensional feature data; for example, this profile data can include user profiles and document profiles, and it can also be multi-dimensional profile data.

[0037] By employing this disclosure, data to be processed can be obtained from multiple data sources, and data processing can be performed on the data to be processed to obtain multi-dimensional feature data corresponding to the data to be processed. The multi-dimensional feature data is used to characterize the various data types corresponding to the data to be processed, and profile data can be constructed based on the multi-dimensional feature data. Thus, profile data describing the multi-dimensional features of the data can be accurately constructed.

[0038] In one embodiment, the method further includes: In one scenario, the data in the first data set is processed based on the profile data to obtain first target data matching the profile data. Alternatively, in another scenario, the data in the second data set is clustered based on the profile data to obtain second target data that has a similarity to the profile data. Using this embodiment, the first scenario is suitable for scenarios requiring precise data identification, such as document download scenarios. In such scenarios, the first target data can be document data that meets user needs. Because the profile data can describe multi-dimensional feature data, the identification accuracy is improved. In the second scenario, the method is suitable for scenarios requiring precise clustering, such as video download scenarios. In such scenarios, the second target data can be video data of interest to similar users in the video download market. Because the profile data can describe multi-dimensional feature data, the clustering accuracy is improved, thereby enabling the precise recommendation of video data of interest to similar users.

[0039] In one embodiment, data processing is performed on the data to be processed to obtain multi-dimensional feature data corresponding to the data to be processed. This includes: performing a first data processing on the data to be processed to obtain a first feature, which is used to characterize the basic features of the first data and / or the first behavior in the data to be processed. Performing a second data processing on the data to be processed to obtain a second feature, which is used to characterize the high-level features in the data to be processed that are associated with the first data and / or the first behavior, and the high-level features associated when the first data and the second data in the data to be processed form interactive data and / or interactive behaviors. Based on the basic features and the high-level features, the multi-dimensional feature data can be obtained. By adopting this embodiment, basic features and high-level data can be obtained respectively through the first data processing (such as statistical analysis in ETL data processing, statistical analysis processing based on statistical results and preset rules) and the second data processing (such as model-based statistical analysis and vector-based statistical analysis processing in ETL data processing). Thus, the multi-dimensional feature data obtained based on the basic features and the high-level features can be used to describe multi-dimensional profile data (such as user profiles and document profiles), which can improve the accuracy of the constructed profile data.

[0040] In one embodiment, constructing profile data based on the multi-dimensional feature data includes: constructing user profiles and document profiles respectively based on the multi-dimensional feature data; and obtaining the profile data based on the user profiles and document profiles. By employing this embodiment, user profiles and document profiles are constructed respectively based on the multi-dimensional feature data obtained from the basic features and the advanced features. Therefore, the profile data obtained from the user profiles and document profiles is also multi-dimensional, improving the accuracy of the constructed profile data.

[0041] In one embodiment, the method further includes: determining the data extracted from the portrait data as initial data, performing third data processing (such as data analysis processing) on ​​the initial data to obtain a third feature. This third feature is used to characterize the feature label corresponding to the initial data. The third feature is then added to the portrait data, and the portrait data is updated. Using this embodiment, taking the portrait data as a portrait feature library (including user portrait features and document portrait features) as an example, to further improve the accuracy of the constructed portrait data, the existing data in the portrait feature library is used as initial data. This initial data (i.e., initial features) is extracted and processed through data analysis to obtain a new third feature (such as a feature label "feature classification label" corresponding to the initial feature in the portrait feature library). This new third feature is then added to the portrait feature library. The updated data obtained by updating the portrait feature library is more accurate and provides better data support.

[0042] In one embodiment, the method further includes: determining the data extracted from the portrait data as initial data; performing a fourth data processing (such as feature engineering analysis) on the initial data to obtain a fourth feature, wherein the fourth feature is used to characterize the filtering features obtained after filtering based on the initial data, and / or derived features related to the initial data. The fourth feature is then added to the portrait data, and the portrait data is updated. Using this embodiment, taking the portrait data as a portrait feature library (including user portrait features and document portrait features) as an example, to further improve the accuracy of the constructed portrait data, the existing data in the portrait feature library is used as initial data. The initial data (i.e., initial features) is extracted and processed through data analysis to obtain a new fourth feature (such as filtering features corresponding to the initial feature in the portrait feature library and / or derived features related to the initial data), and the new fourth feature is added to the portrait feature library. The portrait feature library is then updated, resulting in more accurate updated data and better data support.

[0043] In one embodiment, the method further includes: acquiring target data, which represents the updated data obtained after updating the portrait data, and determining the target data as training data for model training. In this embodiment, the updated portrait data is used as training data for model training; in other words, modeling is performed based on the updated portrait feature library. The resulting model can be used for core member identification in document download scenarios and conversion of existing members, among other business operations.

[0044] The following provides an example illustrating the method for constructing portrait data provided in the embodiments of this disclosure.

[0045] Taking document download scenarios as an example, document profiles and user profiles mainly include some basic features, but lack advanced features. Advanced features can describe the attributes and behaviors of users throughout their entire lifecycle during the profile data construction process. Therefore, it is impossible to accurately construct profile data, and it is also impossible to accurately control the conversion of users at various stages in document download scenarios based on the constructed profile data. For example, it is impossible to identify core members by converting registered members to paid members, or by converting paid members at various levels (such as from basic paid members to intermediate paid members, from monthly members to annual members, etc.).

[0046] The process of users downloading documents involves both basic characteristics (such as user consumption characteristics) and advanced characteristics (such as user production characteristics and user interaction characteristics). These two aspects dynamically change throughout the user's lifecycle, reflecting the attributes and behaviors described at each stage. Therefore, a comprehensive consideration of both basic and advanced characteristics is necessary. Furthermore, considering that the key to attracting users to documents is high-quality document content, and the core of high-quality document content is the existence of producers of high-quality document content (i.e., more users uploading high-quality document content to the document database), and understanding users' document download needs, it's crucial to identify and convert more producers of high-quality document content. This requires multi-dimensional user profiling to accurately classify users, control key conversion nodes throughout the user lifecycle, match multiple business needs, and improve user conversion rates at each stage. This process requires precise data support for all these steps.

[0047] This application example mainly includes two parts: First, based on Figure 3 The illustrated portrait infrastructure provides the basic foundation for portrait creation; secondly, based on... Figure 4 The data update and data modeling process shown realizes the modeling of portrait data.

[0048] I. Basic Construction of the Portrait

[0049] Figure 3 This is a schematic diagram of the image infrastructure in an application example according to an embodiment of this disclosure, such as... Figure 3 As shown, the basic infrastructure of the profile system mainly consists of five layers, from the bottom data source layer to the business application layer: data source, ETL, profile feature library, data storage, and business application.

[0050] (1) Data Source: This integrates data from different data sources, which may include databases, user behavior tables, member tables, authentication data, etc. The database stores all the original data and may include document databases, user behavior data, member data, authentication data, etc. The user behavior table stores all user behaviors, which can be user behaviors in document download scenarios or user behaviors in other business scenarios. The member table stores the relationships between different member levels in the document download scenario. Authentication data stores the mapping between login accounts and other associated accounts. For example, the login account in the document download scenario may be different from the default registered account; authentication is performed through the mapping relationship in this authentication data.

[0051] (2) ETL: Describes data processing modes, including statistical analysis, rule-based statistical analysis, model-based statistical analysis, and vector-based statistical analysis. Statistical analysis can be based on factors such as the number of logins, document views, and document downloads in a document download scenario. Rule-based statistical analysis can be performed after obtaining the statistical results of the statistical analysis, combined with predefined rules. For example, using login frequency as a metric, and with 30 days in a month, the rules can define: logging in for more than 5 days a month qualifies as a moderately active user, logging in for more than 10 days a month qualifies as an active user, etc., i.e., obtaining user activity based on statistical results and rules. Model-based statistical analysis can combine models with business applications to identify core members or existing member conversions, statistically determining the likelihood of core member renewals and the likelihood of existing member conversions and renewals. Vector-based statistical analysis can involve creating a vector each time a user browses a document, calculating the similarity between these vectors, and then using this similarity to describe / identify the document content that the user is interested in. Furthermore, it can determine how many people are similar to each other and are interested in documents with high vector similarity. It's important to note that while rule-based statistical analysis in ETL processing yields these basic features, model-based statistical analysis and vector-based statistical analysis in ETL processing can produce these advanced features.

[0052] (3) Profile Feature Library: The profile feature library consists of two parts: user profiles and document profiles. Within each user profile and document profile, basic features (describing user name, document view count, etc.) and advanced features (different from basic features, these can be features defined according to business needs, such as user activity level, document content quality score, areas of interest in documents, document influence, member conversion rate, and advanced features like "word vectors"). For example... Figure 3As shown, basic features mainly consist of user consumption characteristics, while advanced features mainly consist of user production characteristics and user interaction characteristics. Within each user profile and document profile, various attributes can be included, such as demographic attributes, basic feature attributes, production attributes, consumption attributes, and sequence attributes in the user profile. Among these, production attributes, consumption attributes, and sequence attributes all describe user behavior. Similarly, the document profile includes basic feature attributes, consumption attributes, and sequence attributes. These attributes can be obtained through an ETL process, and based on these attributes, the aforementioned basic features and advanced features can be derived.

[0053] (4) Data storage: This refers to the storage methods for the basic and advanced features mentioned above, and is divided into offline storage and online storage. The offline storage part can be stored on HDFS (HDFS is mainly a distributed storage method for handling massive data storage) for large-scale data retrieval. The UDA tool (UDA is a tool for querying offline data) can be used to query offline data. The online storage part is mainly built on vector libraries (such as the AN vector library) and retrieval libraries (such as the Elasticsearch retrieval library) to provide users with data presentation and provide real-time query and retrieval functions.

[0054] (5) Business Application: The business application can establish a communication connection with the data storage layer through the application programming interface (API). The business application can provide the profile features stored in the data storage layer to the business application party according to different business needs (such as business needs such as member conversion business, document quality needs such as document content scoring, vector distribution needs such as vector similarity calculation based on distribution vectors to determine the documents that users are interested in, profile query needs such as profile feature analysis, user analysis needs such as other user analysis, etc.).

[0055] II. Data Modeling Based on Profile Data

[0056] Figure 4 This is a schematic diagram illustrating image updating and modeling in an application example according to an embodiment of this disclosure, such as... Figure 4 As shown, data modeling based on portrait data mainly includes three parts: data analysis, feature engineering, and modeling based on training data. Among them, data analysis and feature engineering can update the initial portrait feature library to obtain target data for model training, so that the hardware performance of the trained target model (such as classification performance, recognition performance, etc. deployed on electronic devices) is better.

[0057] (1) Data analysis: This involves statistically analyzing the data to be processed stored in the initial profile feature library, obtaining more accurate new features (such as feature labels), and adding these new features to the profile feature library to update the data, thereby obtaining a more accurate multi-dimensional profile feature library. Among these, the new features can be features obtained by classifying users according to business needs and understanding user habits. Based on the new features, model training can be performed more accurately, and entry points for further analysis and modeling can be found.

[0058] (2) Feature engineering: This involves statistically analyzing the data to be processed stored in the initial profile feature library, obtaining more accurate new features (such as filtering features and / or diffraction features), and adding these new features to the profile feature library to update the data, thereby obtaining a more accurate multi-dimensional profile feature library. Among these, the new features can better describe user behavior, and the new features can be used to train the model more accurately.

[0059] (3) Modeling based on training data: Modeling is based on the initial portrait feature library, or updated by any of the above (1)-(2) methods. The model can be a model obtained based on a convolutional neural network. Model selection (lr, randomforest, lightgbm), model optimization (bayesianoptimization), model fusion (bagging, stacking, boosting), etc. can be implemented. Modeling can be implemented according to different business needs. For example, the model can be used to identify core users and convert between members.

[0060] Using this application example, by constructing profile data from the data to be processed, we obtain general multi-dimensional features that describe user needs and document requirements. These multi-dimensional features are stored in the profile feature library, and after data updates, we build the model. This reduces repetitive data work, improves data processing efficiency, and further improves the accuracy of business recommendations by matching different business needs based on the model.

[0061] According to embodiments of this disclosure, an apparatus for constructing portrait data is provided. Figure 5 This is a schematic diagram of the composition structure of the image data construction apparatus according to an embodiment of the present disclosure, such as... Figure 5 As shown, the portrait data construction device 500 includes: a data processing unit 501, used to process data to be processed obtained from multiple data sources to obtain multi-dimensional feature data corresponding to the data to be processed, the multi-dimensional feature data being used to characterize multiple data types corresponding to the data to be processed; and a portrait construction unit 502, used to construct portrait data based on the multi-dimensional feature data.

[0062] In one embodiment, the system further includes: an identification unit, configured to perform identification processing on data in a first data set based on the portrait data to obtain first target data matching the portrait data; or, a clustering unit, configured to perform clustering processing on data in a second data set based on the portrait data to obtain second target data that has a similarity to the portrait data.

[0063] In one embodiment, the data processing unit is configured to perform a first data processing on the data to be processed to obtain a first feature, the first feature being used to characterize the basic features of the first data and / or the first behavior in the data to be processed; perform a second data processing on the data to be processed to obtain a second feature, the second feature being used to characterize the advanced features in the data to be processed that are associated with the first data and / or the first behavior, and the advanced features associated when the first data and the second data in the data to be processed form interactive data and / or interactive behaviors; and obtain the multi-dimensional feature data based on the basic features and the advanced features. For example, the basic features may include consumption features; the advanced features may include production features and / or interaction features.

[0064] In one embodiment, the data processing unit is configured to construct user profiles and document profiles based on the multi-dimensional feature data, and to obtain the profile data based on the user profiles and the document profiles.

[0065] In one embodiment, the system further includes a data update unit, which is used to determine the data extracted from the portrait data as initial data; perform third data processing based on the initial data to obtain a third feature, the third feature being used to characterize the feature label corresponding to the initial data; and add the third feature to the portrait data to update the portrait data.

[0066] In one embodiment, the system further includes a data processing unit, configured to determine the data extracted from the portrait data as initial data; perform a fourth data processing based on the initial data to obtain a fourth feature, the fourth feature being used to characterize the filtering features obtained after filtering based on the initial data, and / or the derived features related to the initial data; and add the fourth feature to the portrait data to update the portrait data.

[0067] In one embodiment, the system further includes a data determination unit for acquiring target data, wherein the target data is used to characterize the updated data obtained after updating the portrait data; and the target data is determined as training data for model training.

[0068] The acquisition, storage, and application of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0069] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0070] Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0071] like Figure 6 As shown, device 600 includes a computing unit 601, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 602 or a computer program loaded into random access memory (RAM) 603 from storage unit 608. RAM 603 may also store various programs and data required for the operation of device 600. The computing unit 601, ROM 602, and RAM 603 are interconnected via bus 604. Input / output (I / O) interface 605 is also connected to bus 604.

[0072] Multiple components in device 600 are connected to I / O interface 605, including: input unit 606, such as keyboard, mouse, etc.; output unit 607, such as various types of monitors, speakers, etc.; storage unit 608, such as disk, optical disk, etc.; and communication unit 609, such as network card, modem, wireless transceiver, etc. Communication unit 609 allows device 600 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0073] The computing unit 601 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 performs the various methods and processes described above, such as the method for constructing portrait data. For example, in some embodiments, the method for constructing portrait data may be implemented as a computer software program tangibly included in a machine-readable medium, such as storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on device 600 via ROM 602 and / or communication unit 609. When the computer program is loaded into RAM 603 and executed by the computing unit 601, one or more steps of the method for constructing portrait data described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the method for constructing portrait data by any other suitable means (e.g., by means of firmware).

[0074] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0075] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0076] In the context of this disclosure, a machine-readable medium can be a tangible medium that may include or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0077] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0078] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0079] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0080] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0081] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for constructing portrait data, comprising: performing first data processing on to-be-processed data obtained from a plurality of data sources to obtain first features, the first features being used to represent basic features of first data and / or first behaviors in the to-be-processed data; performing second data processing on the to-be-processed data to obtain second features, the second features being used to represent advanced features associated with the first data and / or the first behaviors in the to-be-processed data, and advanced features associated with the first data and second data forming interaction data and / or interaction behaviors in the to-be-processed data; obtaining multi-dimensional feature data according to the basic features and the advanced features, wherein the multi-dimensional feature data is used to represent a plurality of data types corresponding to the to-be-processed data, the basic features include consumption features, and the advanced features include production features and interaction features; constructing portrait data according to the multi-dimensional feature data, wherein the portrait data includes user portraits and document portraits, the user portraits include demographic attributes, basic feature attributes, production attributes, consumption attributes, and sequence attributes, the production attributes, the consumption attributes, and the sequence attributes belong to attributes describing user behaviors, and the document portraits include basic feature attributes, consumption attributes, and sequence attributes; performing identification processing on data in a first data set according to the portrait data to obtain first target data matched with the portrait data, wherein the first target data is document data meeting user needs in a document downloading scenario. 2.The method of claim 1, further comprising: performing clustering processing on data in a second data set according to the portrait data to obtain second target data having a similarity with the portrait data.

3. The method of claim 1 or 2, wherein, The constructing portrait data according to the multi-dimensional feature data comprises: constructing user portraits and document portraits respectively according to the multi-dimensional feature data; obtaining the portrait data according to the user portraits and the document portraits. 4.The method of claim 1 or 2, further comprising: determining data extracted from the portrait data as initial data; performing third data processing on the initial data to obtain third features, the third features being used to represent feature labels corresponding to the initial data; adding the third features to the portrait data to perform update processing on the portrait data. 5.The method of claim 1 or 2, further comprising: determining data extracted from the portrait data as initial data; performing fourth data processing on the initial data to obtain fourth features, the fourth features being used to represent screening features obtained based on screening of the initial data, and / or derivative features related to the initial data; adding the fourth features to the portrait data to perform update processing on the portrait data. 6.The method of claim 5, further comprising: obtaining target data, the target data being used to represent update data obtained after performing update processing on the portrait data; determining the target data as training data used for model training. 7.A device for constructing portrait data, comprising: a data processing unit, configured to perform first data processing on to-be-processed data acquired from a plurality of data sources to obtain first features, the first features being used to represent basic features of first data and / or first behaviors in the to-be-processed data; perform second data processing on the to-be-processed data to obtain second features, the second features being used to represent advanced features associated with the first data and / or the first behaviors in the to-be-processed data, and advanced features associated with the first data and second data forming interaction data and / or interaction behaviors in the to-be-processed data; and obtain multi-dimensional feature data according to the basic features and the advanced features; wherein the multi-dimensional feature data is used to represent a plurality of data types corresponding to the to-be-processed data; the basic features include consumption features, and the advanced features include production features and interaction features; a portrait construction unit, configured to construct portrait data according to the multi-dimensional feature data; wherein the portrait data includes user portraits and document portraits; the user portraits include demographic attributes, basic feature attributes, production attributes, consumption attributes, and sequence attributes, the production attributes, the consumption attributes, and the sequence attributes being attributes describing user behaviors; and the document portraits include basic feature attributes, consumption attributes, and sequence attributes; a recognition unit, configured to perform recognition processing on data in a first data set according to the portrait data to obtain first target data matched with the portrait data; wherein the first target data is document data meeting user demands in a document downloading scenario.

8. The apparatus of claim 7, further comprising: a clustering unit, configured to perform clustering processing on data in a second data set according to the portrait data to obtain second target data having a similarity with the portrait data.

9. The apparatus of claim 7 or 8, wherein, The data processing unit is configured to: construct user portraits and document portraits respectively according to the multi-dimensional feature data; and obtain the portrait data according to the user portraits and the document portraits.

10. The apparatus of claim 7 or 8, further comprising a data updating unit, configured to: determine data extracted from the portrait data as initial data; perform third data processing according to the initial data to obtain third features, the third features being used to represent feature labels corresponding to the initial data; and add the third features to the portrait data to perform updating processing on the portrait data.

11. The apparatus of claim 7 or 8, further comprising a data processing unit, configured to: determine data extracted from the portrait data as initial data; perform fourth data processing according to the initial data to obtain fourth features, the fourth features being used to represent filtering features obtained based on filtering of the initial data, and / or derivative features related to the initial data; and add the fourth features to the portrait data to perform updating processing on the portrait data.

12. The apparatus of claim 11, further comprising a data determination unit, configured to: acquire target data, the target data being used to represent updating data obtained after performing updating processing on the portrait data; and determining the target data as training data for model training. 13.An electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

14. A non-transitory computer readable storage medium having stored thereon computer instructions, wherein, the computer instructions are for causing the computer to perform the method of any one of claims 1-6. 15.A computer program product comprising a computer program which, when executed by a processor, implements the method of any one of claims 1-6.

Images