User portrait generation method, device, electronic device and computer storage medium

By identifying and correlating user data of different data types, using neural networks to perform multiple attribute analysis and data fusion, the problem of low user portrait accuracy in the prior art is solved, and user portrait generation with higher accuracy is achieved.

CN113407843BActive Publication Date: 2024-11-15ONE CONNECT SMART TECH CO LTD SHENZHEN
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202110779706.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-07-09
Publication Date
2024-11-15
Estimated Expiration
2041-07-09

AI Technical Summary

Technical Problem

Existing user portrait generation methods cannot effectively associate multiple data sources and modalities, resulting in low accuracy of generated user portraits.

Method used

By identifying the data type of user data, using pre-trained neural networks for content recognition and attribute analysis, a variable subset of user variables is constructed, and data fusion is performed to generate user portraits.

Benefits of technology

The accuracy of user portraits is improved, and the accuracy of analysis results is enhanced through the correlation of multi-source data and the fusion of multiple analysis results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113407843B_ABST
    Figure CN113407843B_ABST
Patent Text Reader

Abstract

The present invention relates to data analysis technology, and discloses a method for generating a user portrait, comprising: performing content identification on the user data according to the data type of the user data to obtain data content; performing a first user attribute analysis on at least two preset user variables using a pre-trained first neural network according to the data content to obtain a first result for each user variable; performing a second user attribute analysis on the user variables using the first result as a parameter of the first neural network to obtain a second result for the user variable; constructing a variable subset of the user variable according to the second result, and performing data fusion on the variable subset using a pre-trained second neural network to obtain a user portrait. In addition, the present invention also relates to blockchain technology, and user data can be stored in a node of a blockchain. The present invention also proposes a user portrait generation device, an electronic device, and a computer-readable storage medium. The present invention can solve the problem of low accuracy when generating user portraits.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to a user portrait generation method, device, electronic device and computer-readable storage medium. Background Art

[0002] Due to the widespread application of user profiling technology in advertising, marketing, risk control, recruitment and other fields, building high-precision user portraits has become increasingly important to Internet companies.

[0003] At present, most methods for constructing user portraits are based on data aggregation, such as storing user data obtained from multiple channels in a unified manner as a user portrait. In this method, the generated user portrait is only a collection of data, and it is impossible to associate user data from multiple sources or modalities (such as text, images, and relational databases), resulting in a low accuracy of the generated user portrait. Summary of the invention

[0004] The present invention provides a user portrait generation method, device and computer-readable storage medium, the main purpose of which is to solve the problem of low accuracy when generating user portraits.

[0005] To achieve the above object, the present invention provides a method for generating a user portrait, comprising:

[0006] Acquire user data, identify the data type of the user data, and perform content identification on the user data using a content identification method corresponding to the data type to obtain data content;

[0007] According to the data content, using a pre-trained first neural network to perform a first user attribute analysis on at least two preset user variables to obtain a first result for each of the user variables;

[0008] Using the first result as a parameter of the first neural network, performing a second user attribute analysis on the user variable to obtain a second result of the user variable;

[0009] A variable subset of the user variable is constructed according to the second result, and a pre-trained second neural network is used to perform data fusion on the variable subset to obtain a user portrait.

[0010] Optionally, identifying the data type of the user data includes: extracting a data type field of each data in the user data; and searching in a preset standard type table according to the data type field to obtain a data type corresponding to the data type field.

[0011] Optionally, before performing a first user attribute analysis on at least two preset user variables using a pre-trained first neural network according to the data content, the method further includes:

[0012] Acquire basic training data, and use the basic training data to perform a first preset number of iterative training on the first neural network to obtain a training result output by the first neural network;

[0013] Obtaining user feedback data on the training results;

[0014] The first neural network is iterated and trained for a second preset number of times using the feedback data and the basic training data to obtain a trained first neural network.

[0015] Optionally, performing a first user attribute analysis on at least two preset user variables using a pre-trained first neural network according to the data content to obtain a first result for each of the user variables includes:

[0016] Using the first neural network to perform vector conversion on the user variable to obtain a user vector, and performing feature extraction and vector conversion on the data content to obtain a content vector;

[0017] Calculating a correlation value between the user vector and the content vector using a preset relationship function;

[0018] Selecting a target vector whose correlation value is greater than a preset threshold from the content vector;

[0019] An activation operation is performed on the target vector and the user vector using a preset activation function to obtain a first result of each of the user variables.

[0020] Optionally, the performing a second user attribute analysis on the user variable using the first result as a parameter of the first neural network to obtain a second result of the user variable includes:

[0021] Selecting target variables from the user variables in sequence;

[0022] Perform parameter conversion on the first results corresponding to the remaining user variables in the user variables except the target variable to obtain result parameters;

[0023] Using the result parameters to assign parameters to the first neural network;

[0024] The target variable is analyzed using the first neural network after parameter assignment to obtain a second result corresponding to the target variable.

[0025] Optionally, constructing a variable subset of the user variable according to the second result includes:

[0026] Gather the second results corresponding to all target variables into a user variable set;

[0027] The second results in the user variable set are grouped multiple times according to different preset numbers, and the results obtained in each grouping are used as the variable subset.

[0028] Optionally, the using a pre-trained second neural network to perform data fusion on the variable subset to obtain a user portrait includes:

[0029] Selecting target subsets from the variable subsets in turn;

[0030] Inputting the second results corresponding to the remaining variable subsets in the variable subsets except the target subset and the data content corresponding to the user variables in the target subset into the second neural network for analysis to obtain a third result;

[0031] Performing vector conversion on the third result to obtain a result vector;

[0032] Each vector in the result vector is concatenated using a preset data aggregation algorithm to obtain a user portrait.

[0033] In order to solve the above problem, the present invention also provides a user portrait generation device, the device comprising:

[0034] A content identification module is used to obtain user data, identify the data type of the user data, and perform content identification on the user data using a content identification method corresponding to the data type to obtain data content;

[0035] A first analysis module, configured to perform a first user attribute analysis on at least two preset user variables using a pre-trained first neural network according to the data content, to obtain a first result for each of the user variables;

[0036] A second analysis module, configured to perform a second user attribute analysis on the user variable by using the first result as a parameter of the first neural network to obtain a second result of the user variable;

[0037] A user portrait generation module is used to construct a variable subset of the user variable according to the second result, and use a pre-trained second neural network to perform data fusion on the variable subset to obtain a user portrait.

[0038] In order to solve the above problem, the present invention further provides an electronic device, the electronic device comprising:

[0039] a memory storing at least one instruction; and

[0040] The processor executes the instructions stored in the memory to implement the user portrait generation method described above.

[0041] In order to solve the above problem, the present invention also provides a computer-readable storage medium, in which at least one instruction is stored. The at least one instruction is executed by a processor in an electronic device to implement the user portrait generation method described above.

[0042] The embodiment of the present invention realizes the generation of user portraits based on multi-source data by identifying user data of different data types, thereby improving the accuracy of user portraits; performs a first user attribute analysis on different user variables, and uses the result value as a parameter to perform a second user attribute analysis on the user variable, thereby improving the accuracy of the analysis result; and constructs a subset of the analysis result, and fuses the analysis result into a user portrait, thereby improving the accuracy of the generated user portrait. Therefore, the user portrait generation method, device, electronic device, and computer-readable storage medium proposed in the present invention can solve the problem of low accuracy of labels generated for resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 A schematic diagram of a flow chart of a method for generating a user portrait according to an embodiment of the present invention;

[0044] Figure 2 A schematic diagram of a training process of a first neural network provided by an embodiment of the present invention;

[0045] Figure 3 A schematic diagram of a process for a second user attribute analysis provided by an embodiment of the present invention;

[0046] Figure 4 A functional module diagram of a user portrait generation device provided by an embodiment of the present invention;

[0047] Figure 5 A schematic diagram of the structure of an electronic device for implementing the user portrait generation method provided in one embodiment of the present invention.

[0048] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0049] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0050] The embodiment of the present application provides a method for generating a user portrait. The execution subject of the user portrait generation method includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the user portrait generation method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0051] Reference Figure 1 FIG. 1 is a flow chart of a method for generating a user portrait according to an embodiment of the present invention. In this embodiment, the method for generating a user portrait includes:

[0052] S1. Acquire user data, identify the data type of the user data, and use a content identification method corresponding to the data type to perform content identification on the user data to obtain data content.

[0053] In the embodiment of the present invention, the user data may be data generated by the user, or data related to the user, such as the user's shopping records, the user's recruitment resume, the user's travel videos, etc.

[0054] The embodiment of the present invention can capture the user data from a pre-built database, network cache or blockchain node for storing the user data.

[0055] In one of the practical application scenarios of the present invention, the user data includes data of multiple data types, for example, photo resumes stored in image form, user videos stored in video form, or user web browsing records stored in data table form.

[0056] In one embodiment of the present invention, identifying the data type of the user data includes: extracting the data type field of each data in the user data; searching in a preset standard type table according to the data type field to obtain the data type corresponding to the data type field.

[0057] In detail, a preset python statement with a data type field extraction function may be used to extract the data type field of each data in the user data, where the data type field is a field in the user data used to identify the data type of each data.

[0058] Specifically, the standard type table includes multiple data type fields and the data type corresponding to each data type field. The standard type table can be uploaded in advance by the user.

[0059] The embodiment of the present invention selects different processing methods according to the data type to perform content identification on the user data.

[0060] In an optional embodiment of the present invention, for the part of the user data in text form, content recognition is performed using technical means with text analysis functions such as NLP (Natural Language Processing) and HMM (Hidden Markov Model) to obtain data content corresponding to the part of the user data in text form; for the part of the user data in image form, content recognition is performed using technical means with image analysis functions such as OCR (Optical Character Recognition) to obtain data content corresponding to the part of the user data in image form; for the part of the user data in video form, content recognition is performed using a combination of technical means with speech analysis functions and image analysis functions such as ASR (Automatic Speech Recognition) and OCR to obtain data content corresponding to the part of the user data in video form.

[0061] For example, the user data includes data 1, data 2, and data 3, where the data type field of data 1 is a text type field, and content recognition is performed using technical means with text analysis functions such as NLP and HMM to obtain the data content corresponding to the text form of the user data; the data type field of data 2 is an image type field, and content recognition is performed using technical means with image analysis functions such as OCR to obtain the data content corresponding to the image form of the user data; the data type field of data 3 is a video type field, and content recognition is performed using a combination of technical means with speech analysis functions and image analysis functions such as ASR and OCR to obtain the data content corresponding to the video form of the user data.

[0062] The embodiment of the present invention performs content recognition on user data according to the data type, which is beneficial to improving the accuracy of obtaining the data content corresponding to each data type.

[0063] S2. Perform a first user attribute analysis on at least two preset user variables using a pre-trained first neural network according to the data content to obtain a first result for each of the user variables.

[0064] In the embodiment of the present invention, the user variable is information related to the user that is needed to generate a user portrait, such as user age, user gender, user occupation, etc. The user variable can be pre-selected by the user.

[0065] In an embodiment of the present invention, a pre-trained first neural network can be used to analyze at least two preset user variables according to the data content to obtain first results corresponding to different user variables. For example, the data content corresponding to the user's shopping record is data content A, and the data content A includes: the items purchased by the user are youth clothing and introductory books on basic computer algorithms; when the preset user variables are the user's age and the user's occupation, the user's age may be 20 to 30 years old through the "youth clothing" in the data content A, and the user's occupation may be an intern programmer through the "introductory books on basic computer algorithms" in the data content A.

[0066] In detail, the first result includes the result about the user variable after analyzing the user variable according to the data content. For example, when the user variable is the user's age, the first result can be the user's age range (such as 20-30 years old), or it can be the estimated value of the user's age (such as 27 years old).

[0067] In one embodiment of the present invention, the first convolutional neural network includes but is not limited to a Bayesian convolutional neural network and a logistic regression convolutional neural network.

[0068] In this embodiment, Figure 2 As shown, the training process of the pre-trained first neural network includes:

[0069] S21, obtaining basic training data, and performing a first preset number of iterative training on the first neural network using the basic training data to obtain a training result output by the first neural network;

[0070] S22, obtaining user feedback data on the training result;

[0071] S23: Perform iterative training on the first neural network for a second preset number of times using the feedback data and the basic training data to obtain a trained first neural network.

[0072] In detail, the first preset number of times and the second preset number of times may be the same.

[0073] Specifically, the training data of the first convolutional neural network includes basic training data and incremental training data. The basic data contains multiple pre-constructed data contents and the true values ​​of user variables corresponding to the data contents; the incremental training data includes inputting the basic training data into the first convolutional neural network for training, obtaining the predicted value of the user variable by the first convolutional neural network, obtaining the user's feedback data on the predicted value, and using these feedback data as incremental training data (for example, a certain data in the incremental training data is the basic training data and the user's feedback data on the predicted value corresponding to the basic training data).

[0074] In this implementation, the user's feedback data is used to correct or confirm the predicted value of the basic training data, thereby optimizing and enriching the training samples to train a more accurate first convolutional neural network and improve the accuracy of the first result obtained by performing the first user attribute analysis on the user variables based on the data content.

[0075] In the embodiment of the present invention, the first user attribute analysis is performed on at least two preset user variables using a pre-trained first neural network according to the data content to obtain a first result for each of the user variables, including:

[0076] Using the first neural network to perform vector conversion on the user variable to obtain a user vector, and performing feature extraction and vector conversion on the data content to obtain a content vector;

[0077] Calculating a correlation value between the user vector and the content vector using a preset relationship function;

[0078] Selecting a target vector whose correlation value is greater than a preset threshold from the content vector;

[0079] An activation operation is performed on the target vector and the user vector using a preset activation function to obtain a first result of each of the user variables.

[0080] In detail, the data content can be pooled, convolved and other operations can be performed through the first neural network to achieve feature extraction of the data content; and the results of feature extraction of the user variables and data content can be converted into vectors using technologies such as word2vec or NLP (Natural Language Processing) to obtain user vectors and content vectors, and then analysis is performed based on the user vectors and content vectors to obtain the first result of each user variable.

[0081] Specifically, the preset relationship function includes a cosine distance function, a Euclidean distance function, etc., and the activation function includes but is not limited to a sigmoid activation function and a softmax activation function.

[0082] For example, there are content vectors corresponding to data content A of a user's shopping record, content vectors corresponding to data content B of a user's web browsing record, and user vectors corresponding to the user's age and user vectors corresponding to the user's occupation. By using a preset relationship function to calculate the association value between the user vector and the content vector, it can be obtained that the association value between the content vector corresponding to data content A and the user vector corresponding to the user's age is 80, the association value between the content vector corresponding to data content A and the user vector corresponding to the user's occupation is 10, the association value between the content vector corresponding to data content B and the user vector corresponding to the user's age is 20, and the association value between the content vector corresponding to data content B and the user vector corresponding to the user's occupation is 90. When the preset threshold is 70, it can be determined that the user vector corresponding to the user's age is the target vector of the content vector corresponding to data content A, and the user vector corresponding to the user's occupation is the target vector of the content vector corresponding to data content B. Furthermore, the activation function is used to perform activation operations on the user vector corresponding to the user's age and the content vector corresponding to data content A to obtain a first result of the user's age (user variable), and the activation function is used to perform activation operations on the user vector corresponding to the user's occupation and the content vector corresponding to data content B to obtain a first result of the user's occupation (user variable).

[0083] S3. Perform a second user attribute analysis on the user variable by using the first result as a parameter of the first neural network to obtain a second result of the user variable.

[0084] In one of the practical application scenarios of the present invention, since the data in the data content all come from the same user, there is a certain correlation between the different data in the data content. Therefore, the embodiment of the present invention uses the first result as a parameter to analyze the user variable to adjust the first result to obtain a more accurate second result.

[0085] In the embodiment of the present invention, Figure 3 As shown, the second user attribute analysis is performed on the user variable using the first result as a parameter of the first neural network to obtain a second result of the user variable, including:

[0086] S31, selecting target variables from the user variables in sequence;

[0087] S32, performing parameter conversion on the first results corresponding to the remaining user variables in the user variables except the target variable to obtain result parameters;

[0088] S33, using the result parameters to assign parameters to the first neural network;

[0089] S34. Analyze the target variable using the first neural network after parameter assignment to obtain a second result corresponding to the target variable.

[0090] In detail, a computer statement with a parameter conversion function can be used to perform parameter conversion on the first result corresponding to the remaining user variables in the user variables except the target variable. For example, the first result corresponding to the remaining user variables in the user variables except the target variable is encoded to obtain a result parameter in the form of encoded data.

[0091] Specifically, the step of analyzing the target variable using the first neural network after parameter assignment is consistent with the step of performing a first user attribute analysis on at least two preset user variables using the pre-trained first neural network in step S2, and will not be described in detail here.

[0092] The second result is a revised value of the first result. For example, the first result includes that the user's age is 20 to 30 years old and the user's occupation is an intern programmer. When the user's age is selected as the target variable, the first result "the user's occupation is an intern programmer" and the data content are input into the first neural network as data content. After analysis, the first result "the user's age is 20 to 30 years old" is revised to the second result: the user's age is 20 to 25 years old.

[0093] The embodiment of the present invention uses the first result as a parameter to perform a second user attribute analysis on the user variable, thereby correcting the first result obtained by the first user attribute analysis to improve the accuracy of the generated second result.

[0094] S4. Construct a variable subset of the user variable according to the second result, and use a pre-trained second neural network to perform data fusion on the variable subset to obtain a user portrait.

[0095] In an embodiment of the present invention, in order to improve the accuracy of the generated user portrait, multiple data contents can be obtained, and each data content will produce second results corresponding to multiple user variables, and then a variable subset containing all user variables is constructed according to the second result generated by each data content, and the variable subset is fused using a pre-trained second neural network to obtain a user portrait.

[0096] For example, if the second result generated by data content A includes: user age q, user occupation w years of work experience as a programmer; the second result generated by data content B includes: user age e, user occupation r years of work experience as a programmer, then the second results corresponding to data content A and data content B are collected and a subset of the second results is generated. For example, subset 1 is (user age q, user occupation w, data content B), subset 2 is (user age e, user occupation w, r years of work experience as a programmer), etc.

[0097] In the embodiment of the present invention, constructing the variable subset of the user variable according to the second result includes:

[0098] Gather the second results corresponding to all target variables into a user variable set;

[0099] The second results in the user variable set are grouped multiple times according to different preset numbers, and the results obtained in each grouping are used as the variable subset.

[0100] In detail, the preset number is usually less than or equal to the number of second results in the number of user variables.

[0101] For example, the user variable set is {a, b, c}, the user variable set is grouped according to the preset number 1 to obtain variable subsets {a}, {b} and {c}, the user variable set is grouped according to the preset number 2 to obtain variable subsets {a, b}, {b, c} and {a, c}, and the user variable set is grouped according to the preset number 3 to obtain the variable subset {a, b, c}.

[0102] In another embodiment of the present invention, a preset subset generation algorithm may be used to operate on the user variable set to generate a variable subset of the user variable. The subset generation algorithm includes but is not limited to an incremental method, a bit vector method, and a binary method.

[0103] Furthermore, an embodiment of the present invention utilizes a pre-trained second neural network to analyze the variable subset to generate a user portrait by combining a second result corresponding to different data contents, wherein the training process of the second neural network is the same as that of the first neural network and is not elaborated here.

[0104] In the embodiment of the present invention, the method of using the pre-trained second neural network to perform data fusion on the variable subset to obtain the user portrait includes:

[0105] Selecting target subsets from the variable subsets in turn;

[0106] Inputting the second results corresponding to the remaining variable subsets in the variable subsets except the target subset and the data content corresponding to the user variables in the target subset into the second neural network for analysis to obtain a third result;

[0107] Performing vector conversion on the third result to obtain a result vector;

[0108] Each vector in the result vector is concatenated using a preset data aggregation algorithm to obtain a user portrait.

[0109] In detail, the step of inputting the second results corresponding to the remaining variable subsets in the variable subsets except the target subset, and the data content corresponding to the user variables in the target subset into the second neural network for analysis is consistent with the step of performing a first user attribute analysis on at least two preset user variables using a pre-trained first neural network in step S2, and will not be repeated here.

[0110] The step of performing vector conversion on the third result is consistent with the step of performing vector conversion on the user variable in step 2, and will not be described in detail here.

[0111] Specifically, the third result is a corrected value of the second result.

[0112] For example, there are variable subsets Y and variable subsets Z, wherein variable subset Y includes a second result generated according to data content A: the user is 20 to 25 years old, and the user's occupation is a programmer with 3 years of work experience; variable subset Z includes a second result generated according to data content B: the user is 20 to 23 years old, and the user's occupation is a programmer with 1 year of work experience; when variable subset Y is selected as the target subset, the second result in the variable subset Z (the user is 20 to 23 years old, and the user's occupation is a programmer with 1 year of work experience) and the data content corresponding to the user variable of the second result in the variable subset Y (the items purchased by the user are youth clothing and introductory books on basic computer algorithms) are input into the second neural network for analysis, and a third result is obtained: the user is 22 years old, and the user's occupation is a programmer with 1.5 years of work experience.

[0113] In this embodiment, the third result can be fused using a preset data aggregation algorithm to obtain the user portrait. The data aggregation algorithm includes but is not limited to: BIRCH algorithm, DBSCAN algorithm, CLRANS algorithm, etc.

[0114] For example, if there is a third result that the user is 22 years old and the user is a programmer with 1.5 years of work experience, the data fusion algorithm is used to convert "the user is 22 years old" into a first vector, and "the user is a programmer with 1.5 years of work experience" into a second vector, and the first vector is combined with the second vector to achieve data fusion of the third result and obtain a user portrait.

[0115] The embodiment of the present invention realizes the generation of user portraits based on multi-source data by identifying user data of different data types, thereby improving the accuracy of user portraits; performs a first user attribute analysis on different user variables, and uses the result value as a parameter to perform a second user attribute analysis on the user variable, thereby improving the accuracy of the analysis result; and constructs a subset of the analysis result, and fuses the analysis result into a user portrait, thereby improving the accuracy of the generated user portrait. Therefore, the user portrait generation method proposed in the present invention can solve the problem of low accuracy of labels generated for resources.

[0116] like Figure 4 , which is a functional module diagram of a user portrait generation device provided by an embodiment of the present invention.

[0117] The user portrait generation device 100 of the present invention can be installed in an electronic device. According to the functions to be implemented, the user portrait generation device 100 can include a content recognition module 101, a first analysis module 102, a second analysis module 103 and a user portrait generation module 104. The module of the present invention can also be referred to as a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0118] In this embodiment, the functions of each module / unit are as follows:

[0119] The content identification module 101 is used to obtain user data, identify the data type of the user data, and perform content identification on the user data using a content identification method corresponding to the data type to obtain data content;

[0120] The first analysis module 102 is used to perform a first user attribute analysis on at least two preset user variables using a pre-trained first neural network according to the data content to obtain a first result for each of the user variables;

[0121] The second analysis module 103 is used to perform a second user attribute analysis on the user variable by using the first result as a parameter of the first neural network to obtain a second result of the user variable;

[0122] The user portrait generation module 104 is used to construct a variable subset of the user variable according to the second result, and use a pre-trained second neural network to perform data fusion on the variable subset to obtain a user portrait.

[0123] In detail, each module described in the user portrait generation device 100 in the embodiment of the present invention is used in the same manner as described above. Figures 1 to 3 The user portrait generation method described in the present invention has the same technical means and can produce the same technical effects, so I will not go into details here.

[0124] like Figure 5 , is a schematic diagram of the structure of an electronic device for implementing a method for generating a user portrait provided by an embodiment of the present invention.

[0125] The electronic device may include a processor 10, a memory 11, a communication bus 12, and a communication interface 13, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a front-end monitoring program.

[0126] In some embodiments, the processor 10 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips. The processor 10 is the control core (ControlUnit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes various functions and processes data of the electronic device by running or executing programs or modules stored in the memory 11 (for example, executing front-end monitoring programs, etc.), and calling data stored in the memory 11.

[0127] The memory 11 includes at least one type of readable storage medium, and the readable storage medium includes a flash memory, a mobile hard disk, a multimedia card, a card-type memory (for example, an SD or DX memory, etc.), a magnetic memory, a disk, an optical disk, etc. In some embodiments, the memory 11 may be an internal storage unit of an electronic device, such as a mobile hard disk of the electronic device. In other embodiments, the memory 11 may also be an external storage device of an electronic device, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Further, the memory 11 may also include both an internal storage unit of the electronic device and an external storage device. The memory 11 may not only be used to store application software and various types of data installed in the electronic device, such as the code of the front-end monitoring program, etc., but may also be used to temporarily store data that has been output or is to be output.

[0128] The communication bus 12 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to realize connection and communication between the memory 11 and at least one processor 10, etc.

[0129] The communication interface 13 is used for communication between the above-mentioned electronic device and other devices, including a network interface and a user interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device and other electronic devices. The user interface may be a display (Display), an input unit (such as a keyboard (Keyboard)), and optionally, the user interface may also be a standard wired interface, a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, etc. Among them, the display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device and to display a visual user interface.

[0130] Figure 5 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 5The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0131] For example, although not shown, the electronic device may also include a power source (such as a battery) for supplying power to various components. Preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include any components such as one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, etc. The electronic device may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0132] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0133] The user portrait generation program stored in the memory 11 of the electronic device 1 is a combination of multiple instructions. When running in the processor 10, it can achieve:

[0134] Acquire user data, identify the data type of the user data, and perform content identification on the user data using a content identification method corresponding to the data type to obtain data content;

[0135] According to the data content, using a pre-trained first neural network to perform a first user attribute analysis on at least two preset user variables to obtain a first result for each of the user variables;

[0136] Using the first result as a parameter of the first neural network, performing a second user attribute analysis on the user variable to obtain a second result of the user variable;

[0137] A variable subset of the user variable is constructed according to the second result, and a pre-trained second neural network is used to perform data fusion on the variable subset to obtain a user portrait.

[0138] Specifically, the specific implementation method of the processor 10 for the above instructions can refer to Figure 1 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0139] Furthermore, if the module / unit integrated in the electronic device 1 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. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0140] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement:

[0141] Acquire user data, identify the data type of the user data, and perform content identification on the user data using a content identification method corresponding to the data type to obtain data content;

[0142] According to the data content, using a pre-trained first neural network to perform a first user attribute analysis on at least two preset user variables to obtain a first result for each of the user variables;

[0143] Using the first result as a parameter of the first neural network, performing a second user attribute analysis on the user variable to obtain a second result of the user variable;

[0144] A variable subset of the user variable is constructed according to the second result, and a pre-trained second neural network is used to perform data fusion on the variable subset to obtain a user portrait.

[0145] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0146] The modules described as separate components may or may not be physically separated, and the components shown as modules 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 modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0147] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, 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 hardware plus software functional modules.

[0148] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0149] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0150] The blockchain referred to in this invention is a new application mode of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0151] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claim can also be implemented by one unit or device through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any particular order.

[0152] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for generating a user portrait, characterized in that: The method comprises: Acquire user data, identify the data type of the user data, and perform content identification on the user data using a content identification method corresponding to the data type to obtain data content; According to the data content, using a pre-trained first neural network to perform a first user attribute analysis on at least two preset user variables to obtain a first result for each of the user variables; Using the first result as a parameter of the first neural network, performing a second user attribute analysis on the user variable to obtain a second result of the user variable; constructing a variable subset of the user variable according to the second result, and performing data fusion on the variable subset using a pre-trained second neural network to obtain a user portrait; Wherein, performing a first user attribute analysis on at least two preset user variables using a pre-trained first neural network according to the data content to obtain a first result for each of the user variables includes: Using the first neural network to perform vector conversion on the user variable to obtain a user vector, and performing feature extraction and vector conversion on the data content to obtain a content vector; Calculating a correlation value between the user vector and the content vector using a preset relationship function; Selecting a target vector whose correlation value is greater than a preset threshold from the content vector; An activation operation is performed on the target vector and the user vector using a preset activation function to obtain a first result of each of the user variables.

2. The user portrait generation method according to claim 1, characterized in that: The identifying the data type of the user data includes: extracting the data type field of each data in the user data; searching in a preset standard type table according to the data type field to obtain the data type corresponding to the data type field.

3. The user portrait generation method according to claim 1, characterized in that: Before performing a first user attribute analysis on at least two preset user variables using a pre-trained first neural network according to the data content, the method further includes: Acquire basic training data, and use the basic training data to perform a first preset number of iterative training on the first neural network to obtain a training result output by the first neural network; Obtaining user feedback data on the training results; The first neural network is iterated and trained for a second preset number of times using the feedback data and the basic training data to obtain a trained first neural network.

4. The user portrait generation method according to claim 1, characterized in that: The step of performing a second user attribute analysis on the user variable using the first result as a parameter of the first neural network to obtain a second result of the user variable includes: Selecting target variables from the user variables in sequence; Perform parameter conversion on the first results corresponding to the remaining user variables in the user variables except the target variable to obtain result parameters; Using the result parameters to assign parameters to the first neural network; The target variable is analyzed using the first neural network after parameter assignment to obtain a second result corresponding to the target variable.

5. The user portrait generation method according to claim 1, characterized in that: The step of constructing a variable subset of the user variable according to the second result includes: Gather the second results corresponding to all target variables into a user variable set; The second results in the user variable set are grouped multiple times according to different preset numbers, and the results obtained in each grouping are used as the variable subset.

6. The user portrait generation method as claimed in claim 1, characterized in that: The using of the pre-trained second neural network to perform data fusion on the variable subset to obtain the user portrait includes: Selecting target subsets from the variable subsets in turn; Inputting the second results corresponding to the remaining variable subsets in the variable subsets except the target subset and the data content corresponding to the user variables in the target subset into the second neural network for analysis to obtain a third result; Performing vector conversion on the third result to obtain a result vector; Each vector in the result vector is concatenated using a preset data aggregation algorithm to obtain a user portrait.

7. A user portrait generation device, used to implement the user portrait generation method according to any one of claims 1 to 6, characterized in that: The device comprises: A content identification module is used to obtain user data, identify the data type of the user data, and perform content identification on the user data using a content identification method corresponding to the data type to obtain data content; A first analysis module, configured to perform a first user attribute analysis on at least two preset user variables using a pre-trained first neural network according to the data content, to obtain a first result for each of the user variables; A second analysis module, configured to perform a second user attribute analysis on the user variable by using the first result as a parameter of the first neural network to obtain a second result of the user variable; A user portrait generation module is used to construct a variable subset of the user variable according to the second result, and use a pre-trained second neural network to perform data fusion on the variable subset to obtain a user portrait.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the user portrait generation method as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the user portrait generation method according to any one of claims 1 to 6 is implemented.