Method, device, equipment and storage medium for identifying category to which user identification belongs

Abstract

The present application relates to a method, device, equipment and storage medium for identifying the category to which a user identification belongs. The method comprises: obtaining characteristic information corresponding to each preset field; generating a factor loading matrix corresponding to each preset field according to the characteristic information corresponding to each preset field; determining the characteristic coefficient of the characteristic information corresponding to each preset field according to the factor loading matrix corresponding to each preset field; determining a resource value according to the characteristic information corresponding to each preset field and the characteristic coefficient of the characteristic information; and determining a category according to the resource value. The method can accurately identify the category.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, apparatus, computer device and storage medium for identifying a category to which a user identification belongs. Background Art

[0002] In traditional technologies, it is often necessary to classify user identifiers so that different actions can be performed according to the categories to which the user identifiers belong. For example, when pushing information to users, corresponding information can be pushed according to the categories to which the user identifiers belong, thereby improving the information conversion rate; or, the risk level of the user identifier can be determined according to the categories to which the user identifiers belong, thereby improving the risk control capability.

[0003] However, due to privacy protection and other reasons, it is sometimes difficult to obtain data directly used for classification. Even if relevant data is used for fitting, the fitted data is mostly inaccurate, which leads to inaccurate classification of user identification. Summary of the invention

[0004] Based on this, it is necessary to provide a user classification method, apparatus, computer device and storage medium that can accurately identify the category to which the user identification belongs in order to solve the above technical problems.

[0005] A user classification method, the method comprising:

[0006] Obtaining feature information corresponding to the target user ID in each preset field;

[0007] According to the characteristic information corresponding to the target user identification in each preset field, a factor loading matrix corresponding to the target user identification in each preset field is generated, and the factor loading matrix is ​​used to represent the correlation between the characteristic information corresponding to the target user identification in each preset field and the target characteristic information corresponding to each preset field;

[0008] Determine the characteristic coefficients of the characteristic information corresponding to the target user identification in each preset field according to the factor loading matrix corresponding to the target user identification in each preset field;

[0009] Determine the resource value of the target user identifier according to the characteristic information corresponding to each preset field of the target user identifier and the characteristic coefficient of the characteristic information;

[0010] The category to which the target user ID belongs is determined according to the resource value of the target user ID.

[0011] A user classification device, the device comprising:

[0012] An acquisition module is used to acquire characteristic information corresponding to the target user identification in each preset field;

[0013] A generation module, used to generate a factor loading matrix corresponding to the target user identifier in each preset field according to the feature information corresponding to the target user identifier in each preset field, wherein the factor loading matrix is ​​used to represent the correlation between the feature information corresponding to the target user identifier in each preset field and the target feature information corresponding to each preset field;

[0014] A determination module, used to determine the characteristic coefficients of the characteristic information corresponding to the target user identification in each preset field according to the factor loading matrix corresponding to the target user identification in each preset field;

[0015] The determination module is further used to determine the resource value of the target user identifier according to the characteristic information corresponding to the target user identifier in each preset field and the characteristic coefficient of the characteristic information;

[0016] The determination module is further used to determine the category to which the target user identifier belongs according to the resource value of the target user identifier.

[0017] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0018] Obtaining feature information corresponding to the target user ID in each preset field;

[0019] According to the characteristic information corresponding to the target user identification in each preset field, a factor loading matrix corresponding to the target user identification in each preset field is generated, and the factor loading matrix is ​​used to represent the correlation between the characteristic information corresponding to the target user identification in each preset field and the target characteristic information corresponding to each preset field;

[0020] Determine the characteristic coefficients of the characteristic information corresponding to the target user identification in each preset field according to the factor loading matrix corresponding to the target user identification in each preset field;

[0021] Determine the resource value of the target user identifier according to the characteristic information corresponding to each preset field of the target user identifier and the characteristic coefficient of the characteristic information;

[0022] The category to which the target user ID belongs is determined according to the resource value of the target user ID.

[0023] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0024] Obtaining feature information corresponding to the target user ID in each preset field;

[0025] According to the characteristic information corresponding to the target user identification in each preset field, a factor loading matrix corresponding to the target user identification in each preset field is generated, and the factor loading matrix is ​​used to represent the correlation between the characteristic information corresponding to the target user identification in each preset field and the target characteristic information corresponding to each preset field;

[0026] Determine the characteristic coefficients of the characteristic information corresponding to the target user identification in each preset field according to the factor loading matrix corresponding to the target user identification in each preset field;

[0027] Determine the resource value of the target user identifier according to the characteristic information corresponding to each preset field of the target user identifier and the characteristic coefficient of the characteristic information;

[0028] The category to which the target user ID belongs is determined according to the resource value of the target user ID.

[0029] The user classification method, device, computer equipment and storage medium obtain the characteristic information corresponding to the target user identification in each preset field, then generate the factor loading matrix corresponding to the target user identification in each preset field according to the characteristic information corresponding to the target user identification in each preset field, then determine the characteristic coefficient of the characteristic information corresponding to the target user identification in each preset field according to the factor loading matrix corresponding to the target user identification in each preset field, then determine the resource value of the target user identification according to the characteristic information corresponding to the target user identification in each preset field and the characteristic coefficient of the characteristic information, and determine the category to which the target user identification belongs according to the resource value of the target user identification, so that the resource value of the target user identification is constructed by comprehensively considering the status of the target user identification in each preset field, and the category to which the target user identification belongs is determined according to the resource value of the target user identification, so as to achieve accurate identification of the category to which the target user identification belongs. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is a flow chart of a user classification method in one embodiment;

[0031] Figure 2 A schematic diagram of a process for constructing a feature identifier of a preset field in an embodiment;

[0032] Figure 3 A schematic diagram of a process for obtaining a factor loading matrix in one embodiment;

[0033] Figure 4 A schematic diagram of a flow chart for obtaining characteristic coefficients in one embodiment;

[0034] Figure 5 is a flow chart of a user classification method in another embodiment;

[0035] Figure 6 is a schematic diagram of an investment factor in one embodiment;

[0036] Figure 7 is a schematic diagram of consumption factors in one embodiment;

[0037] Figure 8 is a schematic diagram of an identity factor in one embodiment;

[0038] Fig. 9 is a schematic diagram of resource factors in one embodiment;

[0039] Fig.10 is a schematic diagram of user asset distribution in one embodiment;

[0040] Fig.11 A schematic diagram of a distribution result of resource values ​​in one embodiment;

[0041] Fig.12 is a structural block diagram of a user classification device in one embodiment;

[0042] Fig.13 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0044] In one embodiment, Figure 1 As shown, a user classification method is provided. This embodiment takes the method applied to a terminal as an example for illustration. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0045] Step 102: Acquire characteristic information corresponding to the target user identification in each preset field.

[0046] The target user identifier refers to the user identifier of the category to be detected. The target user identifier is used to uniquely represent the identity of the target user. The target user identifier can be the target user's ID number, passport number, driver's license number, etc.

[0047] The preset domain may include at least one of an investment domain, a consumption domain, a resource domain, and an identity domain. Investment refers to the process of taking a resource of a preset value as input, processing and outputting a corresponding resource within a preset time through a dynamic adjustment model. Consumption refers to the process of taking a resource of a predetermined value as input, processing and outputting a corresponding resource through an exchange model. Resources may be assets corresponding to the target user identifier, etc. Identity may be a resource level corresponding to the target user identifier.

[0048] The characteristic information in the investment field can be used to describe the investment behavior of the target user identity; the characteristic information in the consumption field can be used to describe the consumption behavior of the target user identity; the characteristic information in the resource field can include the specific assets corresponding to the target user identity; the characteristic information in the identity field can include the factors that constitute the resource level of the target user identity.

[0049] In one embodiment, the feature information may include a feature identifier and a feature value corresponding to the feature identifier of the target user identifier. Obtaining the feature information corresponding to each preset field of the target user identifier includes: obtaining the feature identifier corresponding to each preset field and the feature value corresponding to the feature identifier of the target user identifier.

[0050] In one embodiment, the terminal obtains characteristic information corresponding to the target user identification in each preset field.

[0051] Step 104: Generate a factor loading matrix corresponding to the target user identification in each preset field according to the characteristic information corresponding to the target user identification in each preset field.

[0052] Among them, the factor loading matrix is ​​used to characterize the correlation between the characteristic information corresponding to the target user identification in each preset field and the target characteristic information corresponding to each preset field. The target characteristic information corresponding to each preset field is a latent variable hidden in the characteristic information corresponding to each preset field and cannot be directly measured. For example, taking the clarity of an image as an example, the clarity of an image can be described by variables such as brightness, contrast, saturation, and sharpening intensity. Brightness, contrast, and saturation can be summarized as color factors, and sharpening intensity can be classified as a structural factor. In this way, the four variables are summarized into two latent variables.

[0053] In one embodiment, according to the characteristic information corresponding to the target user identifier in each preset field, the target characteristic information corresponding to the target user identifier in each preset field is determined, and according to the target characteristic information corresponding to the target user identifier in each preset field, the factor loading matrix corresponding to the target user identifier in each preset field is generated. In one embodiment, the factor loading matrix corresponding to the target user identifier in each preset field can be generated according to the characteristic information corresponding to the target user identifier in each preset field by a factor analysis method. The factor analysis method is to find hidden and representative common factors in multiple variables, and classify variables of the same nature into one common factor, thereby reducing the number of variables.

[0054] In one embodiment, a principal component analysis method can be used to generate a factor loading matrix corresponding to the target user identifier in each preset field according to the characteristic information corresponding to the target user identifier in each preset field. The principal component analysis method uses the idea of ​​dimensionality reduction to convert multiple variables into at least one common factor for representation, wherein in the principal component analysis method, the common factor is a linear combination of multiple variables, and each common factor is independent of each other, so that these common factors can reflect most of the information of the variables, and the reflected information does not overlap.

[0055] In one embodiment, the terminal generates a factor loading matrix corresponding to the target user identity in each preset field according to the feature information corresponding to the target user identity in each preset field.

[0056] Step 106, determining the characteristic coefficients of the characteristic information corresponding to the target user identification in each preset field according to the factor loading matrix corresponding to the target user identification in each preset field.

[0057] Among them, the feature coefficient is used to represent the weight of the feature information in the target feature information.

[0058] In one embodiment, the factor loading of the characteristic information corresponding to the target characteristic information can be determined according to the factor loading matrix, and the characteristic coefficient of the characteristic information can be determined according to the factor loading corresponding to the target characteristic information. The factor loading reflects the closeness of the correlation between the i-th variable and the j-th common factor. The larger the absolute value of the factor loading, the closer the correlation between the i-th variable and the j-th common factor.

[0059] In one embodiment, the terminal determines the characteristic coefficient of the characteristic information corresponding to the target user identity in each preset field according to the factor loading matrix corresponding to the target user identity in each preset field.

[0060] Step 108: Determine the resource value of the target user identifier according to the characteristic information corresponding to each preset field of the target user identifier and the characteristic coefficient of the characteristic information.

[0061] The resource value is used to characterize the resource status of the target user ID.

[0062] In one embodiment, the product of the characteristic value corresponding to each preset field of the target user identifier and the characteristic coefficient corresponding to the characteristic value may be calculated, and then the sum of the products may be calculated, and the sum may be used as the resource value of the target user identifier.

[0063] In one embodiment, the terminal determines the resource value of the target user identifier according to the characteristic information corresponding to each preset field of the target user identifier and the characteristic coefficient of the characteristic information.

[0064] Step 110: Determine the category to which the target user identifier belongs according to the resource value of the target user identifier.

[0065] In one embodiment, at least two categories may be preset, such as a first category, a second category, a third category, etc., and an association relationship between pre-stored resource values ​​and categories may be preset. The resource value of the first category may be smaller than the resource value of the third category, and the resource value of the third category may be smaller than the resource value of the second category.

[0066] In one embodiment, a pre-stored association relationship between resource values ​​and categories is obtained, and the category to which the target user identifier belongs is determined according to the resource value of the target user identifier and the association relationship.

[0067] In one embodiment, when information is pushed to a target user ID, the category to which the target user ID belongs may be obtained, and corresponding information may be pushed according to the category to which the target user ID belongs. For example, if the category to which user ID A belongs is the second category, and the category to which user ID B belongs is the first category, information of the investment type may be pushed to user ID A, and information of the financial management type may be pushed to user ID B.

[0068] In one embodiment, when risk identification is performed on a target user identifier, the category to which the target user identifier belongs may be obtained, and the risk level of the target user identifier may be determined based on the category to which the target user identifier belongs. For example, the category to which user identifier A belongs is the second category, and the category to which user identifier B belongs is the first category. When user identifiers A and B apply for resources of the same value, it may be determined that the risk of user identifier A is lower than that of user identifier B.

[0069] In one embodiment, the terminal determines the category to which the target user identifier belongs according to the resource value of the target user identifier.

[0070] In the above user classification method, characteristic information corresponding to the target user identification in each preset field is obtained, and then, based on the characteristic information corresponding to the target user identification in each preset field, a factor loading matrix corresponding to the target user identification in each preset field is generated. Then, based on the factor loading matrix corresponding to the target user identification in each preset field, the characteristic coefficient of the characteristic information corresponding to the target user identification in each preset field is determined. Then, based on the characteristic information corresponding to the target user identification in each preset field and the characteristic coefficient of the characteristic information, the resource value of the target user identification is determined, and the category to which the target user identification belongs is determined based on the resource value of the target user identification. In this way, the resource value of the target user identification is constructed by comprehensively considering the status of the target user identification in each preset field, and the category to which the target user identification belongs is determined based on the asset value of the target user identification, so as to achieve accurate identification of the category to which the target user identification belongs.

[0071] In one embodiment, the method for obtaining the characteristic value corresponding to the characteristic identifier of the target user identifier includes: obtaining the parameter value corresponding to the characteristic identifier of the target user identifier; performing data standardization processing on the parameter value corresponding to the characteristic identifier of the target user identifier to obtain the characteristic value corresponding to the characteristic identifier of the target user identifier.

[0072] The parameter value is used to describe the original data corresponding to the feature identifier.

[0073] In one embodiment, the parameter value may be the original data corresponding to the feature identifier, or may be determined based on the correspondence between the feature identifier and the parameter value. For example, if the feature identifier is "value transfer exceeding the threshold value", the corresponding parameter value may be the original data corresponding to the "value transfer exceeding the threshold value", or the corresponding parameter value may be a value determined based on the pre-stored correspondence between the "value transfer exceeding the threshold value" and the parameter value.

[0074] Specifically, the parameter value corresponding to the feature identifier of the target user identifier is subjected to data standardization processing, so that data of different magnitudes are converted into data of the same magnitude. Optionally, the parameter value corresponding to the feature identifier of the target user identifier can be subjected to data standardization processing by using a Z-Score standardization method, a maximum and minimum standardization method, a decimal calibration method, etc., to obtain the feature value corresponding to the feature identifier of the target user identifier.

[0075] In this embodiment, the parameter value corresponding to the feature identifier of the target user identifier is obtained, and data standardization processing is performed on the parameter value corresponding to the feature identifier of the target user identifier to obtain the feature value corresponding to the feature identifier of the target user identifier. In this way, the feature values ​​corresponding to the feature identifier of the target user identifier are at the same order of magnitude, which is convenient for subsequent analysis and calculation.

[0076] In one embodiment, the method of obtaining the feature identifier corresponding to each preset field includes: obtaining a pre-stored matching relationship between the preset field and the feature identifier; and determining the feature identifier corresponding to each preset field according to the matching relationship.

[0077] In one embodiment, some feature identifiers may be pre-selected, assigned to preset fields according to the attributes of these feature identifiers, and the feature identifiers may be associated with the corresponding preset fields and stored. For example, the feature identifier "whether the value transfer exceeds a threshold" may reflect the consumption habits of the target user identifier, so the feature identifier "whether the value transfer exceeds a threshold" may be associated with the consumption field and stored.

[0078] In this embodiment, a matching relationship between pre-stored preset fields and feature identifiers is obtained, and the feature identifiers corresponding to each preset field are determined based on the matching relationship. In this way, the feature identifiers of each preset field can accurately describe the status of the user in each preset field, thereby improving the accuracy of the numerical analysis of user resources.

[0079] In one embodiment, Figure 2 As shown, the method for constructing the matching relationship between the preset field and the feature identifier includes:

[0080] Step 202: Obtain sample user identifiers corresponding to the first category and sample user identifiers corresponding to the second category, perform feature analysis on the sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category, and determine candidate feature identifiers.

[0081] Among them, with the preset resource value as the boundary, the sample user identifier corresponding to the first category is used to characterize the user type whose resource value is lower than the preset resource value, and the sample user identifier corresponding to the second category is used to characterize the user type whose resource value is higher than the preset resource value. The preset resource value can be set according to actual application.

[0082] In one embodiment, the sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category can be determined based on indicator parameters, the indicator parameters including at least one of the account remaining value, the value transferred to a third-party account, the number of transfers to a personal third-party account, and the number of transfers to other people's accounts, the account remaining value corresponding to the sample user identifiers corresponding to the second category is greater than the account remaining value corresponding to the sample user identifiers corresponding to the first category.

[0083] Optionally, the sample user identifier whose account remaining value, the value transferred to the third-party account, and the number of transfers to the personal third-party account meet the first condition can be used as the sample user identifier corresponding to the first category. When the account remaining value within the preset time period is less than the first preset value, the average value transferred to the third-party account is less than the second preset value, and the average number of transfers to the personal third-party account within the preset time interval is greater than the preset number, it is determined that the account remaining value, the value transferred to the third-party account, and the number of transfers to the personal third-party account meet the first condition. Taking WeChat Wallet as an example, when the remaining value of change in the past 6 months is less than 100 yuan, the average value transferred to the third-party account is less than 10 yuan, and the average number of transfers to the personal third-party account per month is greater than 1 time, it is determined that the account remaining value, the value transferred to the third-party account, and the number of transfers to the personal third-party account meet the first condition. The personal third-party account can be a bank card account number, etc.

[0084] Optionally, the sample user identifiers whose account remaining value and the number of transfers to other people's accounts meet the second condition can be used as sample user identifiers corresponding to the second category. When the average daily account remaining value within the predetermined time period is greater than the second preset value and the average number of transfers to other people's accounts within the predetermined time interval is less than the predetermined number, the account remaining value and the number of transfers to other people's accounts are determined to meet the second condition, wherein the second preset value is greater than the first preset value. Taking WeChat Wallet as an example, when the average daily change remaining value in the past 6 months is greater than 500,000 yuan and the average number of transfers to other people's accounts per month is less than 10 times, the account remaining value and the number of transfers to other people's accounts are determined to meet the second condition, so that the merchant's payment account can be excluded.

[0085] In one embodiment, multiple dimensional portraits may be constructed for the sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category, such as basic portraits, consumption portraits, application preferences, investment preferences, etc., so as to extract multiple candidate feature identifiers.

[0086] Step 204: determine the target feature identifier from the candidate feature identifiers according to the feature selection condition.

[0087] In one embodiment, the feature selection conditions include: at least one of the information value being greater than a first threshold, the recall rate being greater than a second threshold, the statistical period being greater than a third threshold, and the business correlation coefficient being greater than a fourth threshold, and when the feature correlation coefficient of at least two candidate feature identifiers is greater than a fifth threshold, the candidate feature identifier with the highest recall rate is used as the target feature identifier.

[0088] Among them, the information value is the IV (Information Value) value, which is used to describe the predictive ability of the variable. The greater the information value, the stronger the predictive ability of the variable. Optionally, the first threshold value can be 0.2. The recall rate is used to describe how many positive examples in the sample are predicted correctly. Optionally, the second threshold value can be 0.3. The statistical period is used to describe the collection period of the variable data. Optionally, the third threshold value can be 12 months. For example, the candidate feature identifier "the numerical transfer used for consumption in the past month" and the candidate feature identifier "the numerical transfer used for consumption in the past 12 months" are selected as the target feature identifier with higher stability. The business correlation coefficient is used to describe the correlation between the variable and the resource, so that the target feature identifier that is more relevant to the resource is selected, and the candidate feature identifiers such as "page click rate" and "news views" that are not highly relevant to the resource are discarded. Optionally, the fourth threshold value can be 0.7. The feature correlation coefficient is used to describe the correlation between variables. Optionally, the fifth threshold value can be 0.7.

[0089] Step 206, classify the target feature identifiers according to the preset fields, obtain the feature identifiers corresponding to the preset fields, associate the preset fields with the feature identifiers corresponding to the preset fields, and obtain the matching relationship between the preset fields and the feature identifiers.

[0090] In one embodiment, the target feature identifier is assigned to a preset field according to its attributes, and the feature identifier is stored in association with the corresponding preset field. For example, the target feature identifier "whether there is investment experience" can reflect the investment habits of the target user identifier, so the target feature identifier "whether there is investment experience" can be stored in association with the investment field.

[0091] In this embodiment, sample user identifiers corresponding to the first category and sample user identifiers corresponding to the second category are obtained, feature analysis is performed on the sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category, candidate feature identifiers are determined, target feature identifiers are determined from the candidate feature identifiers according to feature selection conditions, the target feature identifiers are classified according to preset fields, feature identifiers corresponding to each preset field are obtained, the preset field is associated with the feature identifiers corresponding to the preset field, and a matching relationship between the preset field and the feature identifier is obtained. In this way, the feature identifier of each preset field can accurately describe the status of the user in each preset field, thereby improving the accuracy of the numerical analysis of user resources.

[0092] In one embodiment, Figure 3 As shown, according to the characteristic information corresponding to the target user identification in each preset field, the factor loading matrix corresponding to the target user identification in each preset field is generated, including:

[0093] Step 302: Generate a correlation coefficient matrix corresponding to each preset field of the target user identification according to the characteristic information corresponding to each preset field of the target user identification.

[0094] The correlation coefficient matrix can reflect the correlation between the feature information, which is established by the correlation coefficient between the feature information, and the correlation coefficient is used to reflect the degree of correlation between the feature information. The larger the correlation coefficient between the feature information, the closer the correlation between the feature information.

[0095] In one embodiment, a covariance matrix corresponding to each preset field of the target user identifier may also be generated according to the feature information corresponding to each preset field of the target user identifier. The covariance matrix may reflect the correlation between the feature information. The greater the covariance between the feature information, the closer the correlation between the feature information.

[0096] Step 304, obtaining the characteristic roots and characteristic vectors of the correlation coefficient matrix corresponding to each preset field.

[0097] In one embodiment, the characteristic root, variance contribution rate and cumulative variance contribution rate of the correlation coefficient matrix corresponding to each preset field are calculated to determine the number of common factors (i.e., target feature information). The characteristic root is used to describe the variance that can be explained by the common factor, the variance contribution rate can reflect the single explanation degree of the common factor to the feature information, and the cumulative variance contribution rate can reflect the cumulative explanation degree of the common factor to the feature information.

[0098] In one embodiment, the number of common factors corresponding to the characteristic roots satisfying the preset conditions (e.g., the characteristic roots are greater than 1) and / or the cumulative variance contribution rate satisfying the predetermined conditions (e.g., the cumulative variance contribution rate is greater than 80%) can be selected in each preset field. Optionally, 1 or 2 common factors can be selected in each preset field.

[0099] Step 306: Generate a factor loading matrix corresponding to the target user identity in each preset field according to the characteristic roots and characteristic vectors of the correlation coefficient matrix corresponding to each preset field.

[0100] In one embodiment, the standard orthogonalized eigenvector corresponding to the characteristic root of the common factor can be calculated, and the factor loading matrix corresponding to the target user identification in each preset field can be established according to the characteristic root of the correlation coefficient matrix and the standard orthogonalized eigenvector of the common factor. Among them, the sum of the squares of the factor loading of each common factor is equal to the corresponding characteristic root. The factor loading is the correlation coefficient between the i-th variable and the j-th common factor, which reflects the closeness of the correlation between the i-th variable and the j-th common factor.

[0101] In this embodiment, based on the characteristic information corresponding to the target user identification in each preset field, a correlation coefficient matrix corresponding to the target user identification in each preset field is generated, and the characteristic roots and characteristic vectors of the correlation coefficient matrix corresponding to each preset field are obtained. Based on the characteristic roots and characteristic vectors of the correlation coefficient matrix corresponding to each preset field, a factor loading matrix corresponding to the target user identification in each preset field is generated. In this way, the factor loading matrix corresponding to the target user identification in each preset field is accurately established.

[0102] In one embodiment, Figure 4 As shown, according to the factor loading matrix corresponding to the target user identification in each preset field, the characteristic coefficients of the characteristic information corresponding to the target user identification in each preset field are determined, including:

[0103] Step 402 , orthogonal rotation is performed on the factor loading matrix corresponding to the target user identification in each preset field to obtain the rotated factor loading matrix corresponding to the target user identification in each preset field.

[0104] Among them, the rotation factor loading matrix is ​​used to characterize the characteristic information corresponding to the target user identification in each preset field, and the correlation between the characteristic information corresponding to each preset field. Specifically, compared with the factor loading matrix, the rotation factor loading matrix separates the factor loading on the same column to 1 and 0, so that each variable generates a larger factor loading on a common factor and a smaller factor loading on other common factors, so that the common factor has a clear meaning. For example, the common factor in the investment field can be the investment factor, the common factor in the consumption field can be the consumption factor, the common factor in the identity field can be the identity factor, and the common factor in the resource field can be the resource factor.

[0105] In one embodiment, the factor loading matrix corresponding to the target user identification in each preset field can be orthogonally rotated by the maximum variance method to obtain the rotated factor loading matrix corresponding to the target user identification in each preset field.

[0106] Step 404: Determine the characteristic coefficients of the characteristic information corresponding to the target user identification in each preset field according to the rotation factor loading matrix.

[0107] In one embodiment, the factor loading of the characteristic information in the target characteristic information can be obtained according to the rotation factor loading matrix, and the characteristic coefficient of the characteristic information corresponding to the target user identification in each preset field can be determined according to the factor loading of the characteristic information in the target characteristic information.

[0108] In this embodiment, an orthogonal rotation process is performed on the factor loading matrix corresponding to the target user identification in each preset field to obtain a rotated factor loading matrix corresponding to the target user identification in each preset field. Based on the rotated factor loading matrix, the characteristic coefficients of the characteristic information corresponding to the target user identification in each preset field are determined. In this way, the characteristic coefficients corresponding to the target user identification in each preset field are accurately obtained.

[0109] In one embodiment, characteristic coefficients of characteristic information corresponding to the target user identification in each preset field are determined based on a rotated factor loading matrix, including: obtaining a factor score coefficient matrix corresponding to the target user identification in each preset field based on the rotated factor loading matrix; and determining the characteristic coefficients of characteristic information corresponding to the target user identification in each preset field based on the factor score coefficient matrix.

[0110] In one embodiment, the rotated factor loading matrix may be processed by weighted least square method, regression analysis prediction method, Thomson factor scoring method, etc. to obtain a factor score coefficient matrix.

[0111] In one embodiment, the characteristic coefficient of the characteristic information may be determined according to the factor score coefficient matrix by obtaining the factor loading of the characteristic information in the target characteristic information and using the factor score coefficient corresponding to the factor loading as the characteristic coefficient of the characteristic information.

[0112] In this embodiment, based on the rotated factor loading matrix, the factor score coefficient matrix corresponding to the target user identification in each preset field is obtained, and based on the factor score coefficient matrix, the characteristic coefficient of the characteristic information corresponding to the target user identification in each preset field is determined. In this way, the characteristic coefficient corresponding to the target user identification in each preset field is accurately obtained.

[0113] In one embodiment, the resource value of the target user identifier is determined based on the characteristic information corresponding to the target user identifier in each preset field and the characteristic coefficient of the characteristic information, including: determining the resource score of the target user identifier in each preset field based on the characteristic information corresponding to the target user identifier in each preset field and the characteristic coefficient of the characteristic information; determining the resource value of the target user identifier based on the resource score of the target user identifier in each preset field.

[0114] In one embodiment, the resource score of the target user identifier in each preset field is determined based on the feature information corresponding to the target user identifier in each preset field and the feature coefficient of the feature information, including: calculating the product between the feature information corresponding to the target user identifier in the preset field and the feature coefficient of the feature information, obtaining the maximum value in the product, and using the maximum value in the product as the resource score of the target user identifier in the preset field; or calculating the product between the feature information corresponding to the target user identifier in the preset field and the feature coefficient of the feature information, calculating the average value of the product, and using the average value of the product as the resource score of the target user identifier in the preset field.

[0115] In one embodiment, determining the resource value of the target user identifier according to the resource score of the target user identifier in each preset field includes:

[0116]

[0117] Among them, y is the resource value of the target user identity, and x is the resource score of the target user identity in each preset field.

[0118] Specifically, log(logarithm) can solve the problem of uneven distribution of resource scores in various preset fields and obtain a normal-like distribution curve; Sigmoid function, that is, f(z) = 1 / (1+e -z ) form, so that the resource value is between 0-100, and when the resource score of a preset field is high, the resource value will also show a relatively high score.

[0119] In one embodiment, the resource value of the target user identifier is determined based on the resource score of the target user identifier in each preset field, including: obtaining the maximum value of the resource scores of each preset field, and using the maximum value as the resource value of the target user identifier; or, calculating the average value of the resource scores of each preset field, and using the average value as the resource value of the target user identifier; or, calculating the sum of the resource scores of each preset field, and using the sum as the resource value of the target user identifier; or, setting weights for the resource scores of each preset field, calculating the product of the resource scores of each preset field and the corresponding weights, and calculating the sum of the products of each preset field to obtain the resource value of the target user identifier.

[0120] In this embodiment, the resource score of the target user identifier in each preset field is determined based on the characteristic information corresponding to the target user identifier in each preset field and the characteristic coefficient of the characteristic information. The resource value of the target user identifier is determined based on the resource score of the target user identifier in each preset field. In this way, the resource value of the target user identifier is constructed through the resource score of the target user identifier in each preset field, which solves the problem of not being able to obtain data directly used to evaluate the user's resource situation.

[0121] In one embodiment, a method for determining a resource score of a target user identifier in a preset field based on feature information corresponding to the target user identifier in the preset field and a feature coefficient of the feature information includes: calculating the product between the feature information corresponding to the target user identifier in the preset field and the feature coefficient of the feature information; and adding the products of the preset fields to obtain the resource score of the target user identifier in the preset field.

[0122] Specifically, the resource score of the target user identifier in the preset field can be calculated according to the following formula:

[0123] F2=∑w i ×F1

[0124] Among them, F2 is the resource score of the preset field, F1 is the feature information corresponding to the target user identifier in the preset field, and w i is the characteristic coefficient of the characteristic information.

[0125] In this embodiment, the product of the characteristic information corresponding to the target user identifier in the preset field and the characteristic coefficient of the characteristic information is calculated, and the products of the preset fields are added to obtain the resource score of the target user identifier in the preset field. In this way, the resource score of the target user identifier in each preset field is obtained, and the resource value of the target user identifier is constructed through the resource score of the target user identifier in each preset field.

[0126] In one embodiment, Figure 5 As shown, a user classification method is provided, the method comprising the following steps:

[0127] Step 502: Obtain sample user identifiers corresponding to the first category and sample user identifiers corresponding to the second category, perform feature analysis on the sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category, and determine candidate feature identifiers.

[0128] Step 504: determine the target feature identifier from the candidate feature identifiers according to the feature selection condition.

[0129] Step 506: classify the target feature identifiers according to the preset fields to obtain feature identifiers corresponding to the preset fields.

[0130] Step 508, obtaining the feature identifiers corresponding to the preset fields and the feature values ​​corresponding to the target user identifiers in the feature identifiers, the preset fields include investment field, consumption field, resource field and identity field.

[0131] Step 510: Determine target feature information corresponding to the target user identifier in each preset field according to the feature identifier corresponding to each preset field and the feature value corresponding to the target user identifier in the feature identifier.

[0132] Step 512: determining the characteristic coefficient of the characteristic information corresponding to the target user identifier in each preset field according to the target characteristic information corresponding to the target user identifier in each preset field.

[0133] Step 514, determining the resource value corresponding to the target user identity according to the characteristic information corresponding to each preset field of the target user identity and the characteristic coefficient of the characteristic information, and determining the category to which the target user identity belongs according to the resource value corresponding to the target user identity.

[0134] In this embodiment, the resource value of the target user identity is constructed based on the status of the target user identity in various preset fields, so as to accurately analyze the resource situation of the user identity, and then determine the category to which the target user identity belongs based on the asset value of the target user identity, so as to accurately identify the category to which the target user identity belongs.

[0135] In a specific embodiment, a user classification method is provided, the method comprising the following steps:

[0136] Acquire feature information corresponding to the target user identifier in each preset field. The preset fields may include investment field, consumption field, resource field and identity field. The feature information may include a feature identifier and a feature value corresponding to the target user identifier in the feature identifier.

[0137] Next, based on the characteristic information corresponding to the target user identification in each preset field, a correlation coefficient matrix corresponding to the target user identification in each preset field is generated, and the characteristic roots and eigenvectors of the correlation coefficient matrix corresponding to each preset field are obtained. Based on the characteristic roots and eigenvectors of the correlation coefficient matrix corresponding to each preset field, a factor loading matrix corresponding to the target user identification in each preset field is generated.

[0138] Furthermore, an orthogonal rotation process is performed on the factor loading matrix corresponding to the target user identification in each preset field to obtain a rotated factor loading matrix corresponding to the target user identification in each preset field.

[0139] Next, based on the rotated factor loading matrix, the factor score coefficient matrix corresponding to the target user identification in each preset field is obtained, and based on the factor score coefficient matrix, the characteristic coefficient of the characteristic information corresponding to the target user identification in each preset field is determined. Figure 6 As shown, Figure 6 It shows the characteristic coefficients between the characteristic information of the investment field and the common factors; Figure 7 As shown, Figure 7 The characteristic coefficients between the characteristic information of the consumption field and the common factors are shown; Figure 8 As shown, Figure 8 The characteristic coefficients between the characteristic information of the identity domain and the common factors are shown; e.g. Fig. 9 As shown, Fig. 9 The characteristic coefficients between the characteristic information of the resource field and the common factors are displayed.

[0140] Furthermore, the resource score of the target user identifier in each preset field is determined according to the characteristic information corresponding to the target user identifier in each preset field and the characteristic coefficient of the characteristic information.

[0141] Next, the resource value of the target user identifier is determined according to the resource score of the target user identifier in each preset field.

[0142] Specifically, the resource value of the target user identifier can be calculated by the following formula.

[0143]

[0144] Among them, y is the resource value of the target user identity, and x is the resource score of the target user identity in each preset field.

[0145] Furthermore, the category to which the target user identifier belongs is determined according to the resource value of the target user identifier.

[0146] like Fig.10 As shown, Fig.10 The user asset distribution predicted by the method provided in this embodiment is consistent with the data provided in the research report.

[0147] like Fig.11 As shown, Fig.11 This is a comparison of the resource value distribution of high net worth users and general net worth users. The left side shows the proportion of general net worth users, and the right side shows the proportion of high net worth users. The resource values ​​of the two types of users are clearly distinguished, indicating that the resource values ​​can accurately reflect the user's resource situation.

[0148] In the above user classification method, the resource value of the target user identity is constructed by comprehensively considering the status of the target user identity in various preset fields, so as to accurately analyze the resource situation of the user identity, and then determine the category to which the target user identity belongs according to the asset value of the target user identity, so as to accurately identify the category to which the target user identity belongs.

[0149] The present application also provides an application scenario, which applies the above user classification method. Specifically, the user classification method is applied in the application scenario as follows: analyzing the category to which the target user identifier belongs, and pushing information to the target user identifier according to the category to which the target user identifier belongs.

[0150] The method includes:

[0151] Acquire feature information corresponding to the target user identifier in each preset field. The preset fields may include investment field, consumption field, resource field and identity field. The feature information may include a feature identifier and a feature value corresponding to the target user identifier in the feature identifier.

[0152] Next, based on the characteristic information corresponding to the target user identification in each preset field, a correlation coefficient matrix corresponding to the target user identification in each preset field is generated, and the characteristic roots and eigenvectors of the correlation coefficient matrix corresponding to each preset field are obtained. Based on the characteristic roots and eigenvectors of the correlation coefficient matrix corresponding to each preset field, a factor loading matrix corresponding to the target user identification in each preset field is generated.

[0153] Furthermore, an orthogonal rotation process is performed on the factor loading matrix corresponding to the target user identification in each preset field to obtain a rotated factor loading matrix corresponding to the target user identification in each preset field.

[0154] Next, based on the rotated factor loading matrix, the factor score coefficient matrix corresponding to the target user identification in each preset field is obtained, and based on the factor score coefficient matrix, the characteristic coefficient of the characteristic information corresponding to the target user identification in each preset field is determined.

[0155] Furthermore, the resource score of the target user identifier in each preset field is determined according to the characteristic information corresponding to the target user identifier in each preset field and the characteristic coefficient of the characteristic information.

[0156] Next, the resource value of the target user identifier is determined according to the resource score of the target user identifier in each preset field.

[0157] Specifically, the resource value of the target user identifier can be calculated by the following formula.

[0158]

[0159] Among them, y is the resource value of the target user identity, and x is the resource score of the target user identity in each preset field.

[0160] Furthermore, the category to which the target user identifier belongs is determined according to the resource value of the target user identifier.

[0161] Next, information is pushed to the target user ID according to the category to which the target user ID belongs. An association relationship between the category and the information type can be preset, and information of the corresponding information type is pushed to the target user ID according to the category to which the target user ID belongs.

[0162] In the above user classification method, the resource value of the target user identity is constructed by comprehensively considering the status of the target user identity in various preset fields, so as to accurately analyze the resource situation of the user identity, and then determine the category to which the target user identity belongs according to the asset value of the target user identity, so as to accurately identify the category to which the target user identity belongs, thereby pushing information to the target user identity in a targeted manner, thereby improving the information conversion rate.

[0163] The present application also provides an application scenario, which applies the above user classification method. Specifically, the user classification method is applied in the application scenario as follows: analyzing the category to which the target user identifier belongs, and performing risk identification on the target user identifier according to the category to which the target user identifier belongs.

[0164] The method includes:

[0165] Acquire feature information corresponding to the target user identifier in each preset field. The preset fields may include investment field, consumption field, resource field and identity field. The feature information may include a feature identifier and a feature value corresponding to the target user identifier in the feature identifier.

[0166] Next, based on the characteristic information corresponding to the target user identification in each preset field, a correlation coefficient matrix corresponding to the target user identification in each preset field is generated, and the characteristic roots and eigenvectors of the correlation coefficient matrix corresponding to each preset field are obtained. Based on the characteristic roots and eigenvectors of the correlation coefficient matrix corresponding to each preset field, a factor loading matrix corresponding to the target user identification in each preset field is generated.

[0167] Furthermore, an orthogonal rotation process is performed on the factor loading matrix corresponding to the target user identification in each preset field to obtain a rotated factor loading matrix corresponding to the target user identification in each preset field.

[0168] Next, based on the rotated factor loading matrix, the factor score coefficient matrix corresponding to the target user identification in each preset field is obtained, and based on the factor score coefficient matrix, the characteristic coefficient of the characteristic information corresponding to the target user identification in each preset field is determined.

[0169] Furthermore, the resource score of the target user identifier in each preset field is determined according to the characteristic information corresponding to the target user identifier in each preset field and the characteristic coefficient of the characteristic information.

[0170] Next, the resource value of the target user identifier is determined according to the resource score of the target user identifier in each preset field.

[0171] Specifically, the resource value of the target user identifier can be calculated by the following formula.

[0172]

[0173] Among them, y is the resource value of the target user identity, and x is the resource score of the target user identity in each preset field.

[0174] Furthermore, the category to which the target user identifier belongs is determined according to the resource value of the target user identifier.

[0175] Next, the risk level of the target user identification is detected according to the category to which the target user identification belongs. The association relationship between the category and the risk level can be preset, and the risk level of the target user identification is determined according to the category to which the target user identification belongs.

[0176] In the above user classification method, the resource value of the target user identity is constructed by comprehensively considering the status of the target user identity in various preset fields, so as to accurately analyze the resource situation of the user identity, and then determine the category to which the target user identity belongs according to the asset value of the target user identity, so as to accurately identify the category to which the target user identity belongs, thereby determining the risk level of the target user identity and improving the risk control capability.

[0177] It should be understood that although Figure 1-5 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1-5 At least part of the steps may include multiple steps or multiple stages. These steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.

[0178] In one embodiment, Fig.12As shown, a user classification device is provided, which can be a software module or a hardware module, or a combination of the two to become a part of a computer device. The device specifically includes: an acquisition module 1202, a generation module 1204 and a determination module 1206, wherein:

[0179] An acquisition module 1202 is used to acquire characteristic information corresponding to the target user identification in each preset field;

[0180] A generating module 1204 is used to generate a factor loading matrix corresponding to each preset field of the target user identifier according to the feature information corresponding to each preset field of the target user identifier, wherein the factor loading matrix is ​​used to represent the correlation between the feature information corresponding to each preset field of the target user identifier and the target feature information corresponding to each preset field;

[0181] A determination module 1206 is used to determine the characteristic coefficients of the characteristic information corresponding to the target user identification in each preset field according to the factor loading matrix corresponding to the target user identification in each preset field;

[0182] The determination module 1206 is further used to determine the resource value of the target user identifier according to the characteristic information corresponding to each preset field of the target user identifier and the characteristic coefficient of the characteristic information;

[0183] The determination module 1206 is further configured to determine the category to which the target user identifier belongs according to the resource value of the target user identifier.

[0184] In the above-mentioned user classification device, sample user identifiers corresponding to the first category and sample user identifiers corresponding to the second category are obtained, feature analysis is performed on the sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category to determine candidate feature identifiers, and according to feature selection conditions, target feature identifiers are determined from the candidate feature identifiers, and the target feature identifiers are classified according to preset fields to obtain feature identifiers corresponding to each preset field, and the preset field is associated with the feature identifiers corresponding to the preset field to obtain a matching relationship between the preset field and the feature identifier. In this way, the resource value of the target user identifier is constructed by comprehensively considering the status of the target user identifier in each preset field, and the category to which the target user identifier belongs is determined according to the asset value of the target user identifier, so as to achieve accurate identification of the category to which the target user identifier belongs.

[0185] In one embodiment, the feature information includes a feature identifier and a feature value corresponding to the feature identifier of the target user; the acquisition module 1202 is also used to: obtain the feature identifier corresponding to each preset field, and the feature value corresponding to the feature identifier of the target user; the acquisition module 1202 is also used to: obtain the parameter value corresponding to the feature identifier of the target user; perform data standardization processing on the parameter value corresponding to the feature identifier of the target user to obtain the feature value corresponding to the feature identifier of the target user.

[0186] In this embodiment, the parameter value corresponding to the feature identifier of the target user identifier is obtained, and data standardization processing is performed on the parameter value corresponding to the feature identifier of the target user identifier to obtain the feature value corresponding to the feature identifier of the target user identifier. In this way, the feature values ​​corresponding to the feature identifier of the target user identifier are at the same order of magnitude, which is convenient for subsequent analysis and calculation.

[0187] In one embodiment, the acquisition module 1202 is further used to: acquire a pre-stored matching relationship between preset fields and feature identifiers; and determine the feature identifier corresponding to each preset field according to the matching relationship.

[0188] In this embodiment, a matching relationship between pre-stored preset fields and feature identifiers is obtained, and the feature identifiers corresponding to each preset field are determined based on the matching relationship. In this way, the feature identifiers of each preset field can accurately describe the status of the user in each preset field, thereby improving the accuracy of the numerical analysis of user resources.

[0189] In one embodiment, the acquisition module 1202 is further used to: acquire the sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category, perform feature analysis on the sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category, and determine candidate feature identifiers; the determination module 1206 is further used to: determine the target feature identifier from the candidate feature identifiers according to the feature selection conditions; the user classification device also includes a classification module, the classification module is used to: classify the target feature identifiers according to the preset fields, obtain the feature identifiers corresponding to each preset field, associate the preset field with the feature identifiers corresponding to the preset field, and obtain the matching relationship between the preset field and the feature identifier; wherein, the first category corresponds to The sample user identifier of the first category and the sample user identifier corresponding to the second category are determined according to indicator parameters, and the indicator parameters include at least one of the account remaining value, the value transferred to the third-party account, the number of transfers to the personal third-party account and the number of transfers to other people's accounts. The account remaining value corresponding to the sample user identifier corresponding to the second category is greater than the account remaining value corresponding to the sample user identifier corresponding to the first category; the feature selection conditions include: at least one of the information value is greater than the first threshold, the recall rate is greater than the second threshold, the statistical period is greater than the third threshold, and the business correlation coefficient is greater than the fourth threshold, and when the feature correlation coefficient of at least two candidate feature identifiers is greater than the fifth threshold, the candidate feature identifier with the highest recall rate is used as the target feature identifier.

[0190] In this embodiment, sample user identifiers corresponding to the first category and sample user identifiers corresponding to the second category are obtained, feature analysis is performed on the sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category, candidate feature identifiers are determined, target feature identifiers are determined from the candidate feature identifiers according to feature selection conditions, the target feature identifiers are classified according to preset fields, feature identifiers corresponding to each preset field are obtained, the preset field is associated with the feature identifiers corresponding to the preset field, and a matching relationship between the preset field and the feature identifier is obtained. In this way, the feature identifier of each preset field can accurately describe the status of the user in each preset field, thereby improving the accuracy of the numerical analysis of user resources.

[0191] In one embodiment, the generation module 1204 is also used to: generate a correlation coefficient matrix corresponding to the target user identity in each preset field according to the characteristic information corresponding to the target user identity in each preset field; obtain the characteristic roots and characteristic vectors of the correlation coefficient matrix corresponding to each preset field; generate a factor loading matrix corresponding to the target user identity in each preset field according to the characteristic roots and characteristic vectors of the correlation coefficient matrix corresponding to each preset field.

[0192] In this embodiment, based on the characteristic information corresponding to the target user identification in each preset field, a correlation coefficient matrix corresponding to the target user identification in each preset field is generated, and the characteristic roots and characteristic vectors of the correlation coefficient matrix corresponding to each preset field are obtained. Based on the characteristic roots and characteristic vectors of the correlation coefficient matrix corresponding to each preset field, a factor loading matrix corresponding to the target user identification in each preset field is generated. In this way, the factor loading matrix corresponding to the target user identification in each preset field is accurately established.

[0193] In one embodiment, the determination module 1206 is further used to: perform orthogonal rotation processing on the factor loading matrix corresponding to the target user identification in each preset field to obtain the rotated factor loading matrix corresponding to the target user identification in each preset field, the rotated factor loading matrix is ​​used to represent the characteristic information corresponding to the target user identification in each preset field, and the correlation between the characteristic information and the target characteristic information corresponding to each preset field; determine the characteristic coefficient of the characteristic information corresponding to the target user identification in each preset field according to the rotated factor loading matrix.

[0194] In this embodiment, an orthogonal rotation process is performed on the factor loading matrix corresponding to the target user identification in each preset field to obtain a rotated factor loading matrix corresponding to the target user identification in each preset field. Based on the rotated factor loading matrix, the characteristic coefficients of the characteristic information corresponding to the target user identification in each preset field are determined. In this way, the characteristic coefficients corresponding to the target user identification in each preset field are accurately obtained.

[0195] In one embodiment, the determination module 1206 is further used to: obtain the factor score coefficient matrix corresponding to the target user identification in each preset field according to the rotated factor loading matrix; and determine the characteristic coefficient of the characteristic information corresponding to the target user identification in each preset field according to the factor score coefficient matrix.

[0196] In this embodiment, based on the rotated factor loading matrix, the factor score coefficient matrix corresponding to the target user identification in each preset field is obtained, and based on the factor score coefficient matrix, the characteristic coefficient of the characteristic information corresponding to the target user identification in each preset field is determined. In this way, the characteristic coefficient corresponding to the target user identification in each preset field is accurately obtained.

[0197] In one embodiment, the determination module 1206 is further used to: determine the resource score of the target user identifier in each preset field based on the characteristic information corresponding to the target user identifier in each preset field, and the characteristic coefficient of the characteristic information; determine the resource value of the target user identifier based on the resource score of the target user identifier in each preset field.

[0198] In this embodiment, the resource score of the target user identifier in each preset field is determined based on the characteristic information corresponding to the target user identifier in each preset field and the characteristic coefficient of the characteristic information. The resource value of the target user identifier is determined based on the resource score of the target user identifier in each preset field. In this way, the resource value of the target user identifier is constructed through the resource score of the target user identifier in each preset field, which solves the problem of not being able to obtain data directly used to evaluate the user's resource situation.

[0199] In one embodiment, the determination module 1206 is further used to: calculate the product of the characteristic information corresponding to the target user identifier in the preset field and the characteristic coefficient of the characteristic information; and add the products of the preset fields to obtain the resource score of the target user identifier in the preset field.

[0200] In this embodiment, the product of the characteristic information corresponding to the target user identifier in the preset field and the characteristic coefficient of the characteristic information is calculated, and the products of the preset fields are added to obtain the resource score of the target user identifier in the preset field. In this way, the resource score of the target user identifier in each preset field is obtained, and the resource value of the target user identifier is constructed through the resource score of the target user identifier in each preset field.

[0201] For the specific definition of the user classification device, please refer to the definition of the user classification method above, which will not be repeated here. Each module in the above user classification device can be implemented in whole or in part by software, hardware and a combination thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0202] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Fig.13As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a user classification method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0203] Those skilled in the art will understand that Fig.13 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0204] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0205] In one embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0206] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0207] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0208] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for identifying a category to which a user identification belongs, characterized in that: The method comprises: Acquire a pre-stored matching relationship between a preset field and a feature identifier, and acquire feature information corresponding to the target user identifier in each preset field according to the matching relationship, wherein the feature information includes the feature identifier and a feature value corresponding to the target user identifier in the feature identifier, and the preset fields include an investment field, a consumption field, a resource field, and an identity field; According to the characteristic information corresponding to each of the preset fields, the factor loading matrix corresponding to each of the preset fields of the target user identifier is generated, wherein the factor loading matrix is ​​used to represent the correlation between the characteristic information corresponding to each of the preset fields of the target user identifier and the target characteristic information corresponding to each of the preset fields; Performing orthogonal rotation processing on the factor loading matrix corresponding to each of the preset fields of the target user identification, obtaining the rotated factor loading matrix corresponding to each of the preset fields of the target user identification, the rotated factor loading matrix being used to characterize the correlation between the characteristic information corresponding to each of the preset fields of the target user identification and the target characteristic information corresponding to each of the preset fields; obtaining the factor score coefficient matrix corresponding to each of the preset fields of the target user identification according to the rotated factor loading matrix; determining the characteristic coefficient of the characteristic information corresponding to each of the preset fields of the target user identification according to the factor score coefficient matrix, the characteristic coefficient being used to characterize the weight of the characteristic information in the target characteristic information; Determine the resource value corresponding to the target user identifier according to the characteristic information corresponding to each of the preset fields and the characteristic coefficient of the characteristic information; Determine the category to which the target user identifier belongs according to the resource value corresponding to the target user identifier; Pushing corresponding information to the target user identifier according to the category to which the target user identifier belongs; The method of constructing the matching relationship between the preset domain and the feature identifier includes: The sample user identifiers whose account remaining value, the value transferred to the third-party account, and the number of times the value is transferred to the personal third-party account meet the first condition are used as sample user identifiers corresponding to the first category, and the sample user identifiers whose account remaining value and the number of times the value is transferred to other people's accounts meet the second condition are used as sample user identifiers corresponding to the second category, and the account remaining value corresponding to the sample user identifier corresponding to the second category is greater than the account remaining value corresponding to the sample user identifier corresponding to the first category; feature analysis is performed on the sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category to determine candidate feature identifiers; Determining a target feature identifier from among the candidate feature identifiers according to a feature selection condition; The target feature identifiers are classified according to the preset fields to obtain feature identifiers corresponding to the preset fields, and the preset fields are associated with the feature identifiers corresponding to the preset fields to obtain a matching relationship between the preset fields and the feature identifiers.

2. The method according to claim 1, characterized in that: The method of obtaining the characteristic value corresponding to the characteristic identifier of the target user identifier includes: Obtaining a parameter value of the target user identifier corresponding to the feature identifier; Data standardization is performed on the parameter value of the target user identifier corresponding to the feature identifier to obtain the feature value of the target user identifier corresponding to the feature identifier.

3. The method according to claim 2, characterized in that The sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category are determined based on indicator parameters, and the indicator parameters include at least one of the account remaining value, the value transferred to the personal third-party account, the number of transfers to the personal third-party account, and the number of transfers to other people's accounts.

4. The method according to claim 2, characterized in that: The feature selection conditions include: at least one of the information value being greater than a first threshold, the recall rate being greater than a second threshold, the statistical period being greater than a third threshold, and the business correlation coefficient being greater than a fourth threshold, and when the feature correlation coefficient of at least two of the candidate feature identifiers is greater than a fifth threshold, the candidate feature identifier with the highest recall rate is used as the target feature identifier.

5. The method according to claim 1, characterized in that The step of generating a factor loading matrix corresponding to each of the preset fields according to the characteristic information corresponding to each of the preset fields of the target user identifier includes: Generate a correlation coefficient matrix corresponding to each of the preset fields according to the characteristic information corresponding to each of the preset fields of the target user identifier; Obtaining the characteristic roots and characteristic vectors of the correlation coefficient matrix corresponding to each of the preset fields; According to the characteristic roots and characteristic vectors of the correlation coefficient matrix corresponding to each of the preset fields, a factor loading matrix corresponding to the target user identity in each of the preset fields is generated.

6. The method according to claim 1, characterized in that The determining of the resource value corresponding to the target user identifier according to the characteristic information corresponding to each of the preset fields and the characteristic coefficient of the characteristic information includes: Determine the resource score of the target user identifier in each of the preset fields according to the feature information corresponding to the target user identifier in each of the preset fields and the feature coefficient of the feature information; The resource value of the target user identifier is determined according to the resource score of the target user identifier in each of the preset fields.

7. The method according to claim 6, characterized in that The method of determining the resource score of the target user identifier in the preset field according to the feature information corresponding to the target user identifier in the preset field and the feature coefficient of the feature information includes: Calculate the product of the characteristic information corresponding to the target user identifier in the preset field and the characteristic coefficient of the characteristic information; The products of the preset fields are summed up to obtain the resource score of the target user identifier in the preset field.

8. A device for identifying the category to which a user identification belongs, characterized in that: The device comprises: An acquisition module, used to acquire a pre-stored matching relationship between a preset field and a feature identifier, and acquire feature information corresponding to the target user identifier in each preset field according to the matching relationship, wherein the feature information includes the feature identifier and a feature value corresponding to the target user identifier in the feature identifier, and the preset fields include an investment field, a consumption field, a resource field, and an identity field; A generating module, used to generate a factor loading matrix corresponding to each of the preset fields according to the characteristic information corresponding to each of the preset fields of the target user identifier, wherein the factor loading matrix is ​​used to represent the correlation between the characteristic information corresponding to each of the preset fields of the target user identifier and the target characteristic information corresponding to each of the preset fields; A determination module is used to perform orthogonal rotation processing on the factor load matrix corresponding to the target user identifier in each of the preset fields to obtain the rotated factor load matrix corresponding to the target user identifier in each of the preset fields, and the rotated factor load matrix is ​​used to characterize the correlation between the feature information corresponding to the target user identifier in each of the preset fields and the target feature information corresponding to each of the preset fields; according to the rotated factor load matrix, obtain the factor score coefficient matrix corresponding to the target user identifier in each of the preset fields; according to the factor score coefficient matrix, determine the feature coefficient of the feature information corresponding to the target user identifier in each of the preset fields, and the feature coefficient is used to characterize the weight of the feature information in the target feature information; The determination module is further configured to determine the resource value of the target user identifier according to the characteristic information corresponding to each of the preset fields and the characteristic coefficient of the characteristic information; The determination module is further used to determine the category to which the target user identifier belongs according to the resource value of the target user identifier, and push corresponding information to the target user identifier according to the category to which the target user identifier belongs; Among them, the acquisition module is also used to use the sample user identifiers whose account remaining value, the value transferred to the third-party account, and the number of transfers to the personal third-party account meet the first condition as the sample user identifiers corresponding to the first category, and use the sample user identifiers whose account remaining value and the number of transfers to other people's accounts meet the second condition as the sample user identifiers corresponding to the second category, and the account remaining value corresponding to the sample user identifier corresponding to the second category is greater than the account remaining value corresponding to the sample user identifier corresponding to the first category; perform feature analysis on the sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category to determine candidate feature identifiers; determine a target feature identifier from the candidate feature identifiers according to the feature selection condition; classify the target feature identifiers according to the preset fields to obtain the feature identifiers corresponding to each of the preset fields, associate the preset fields with the feature identifiers corresponding to the preset fields, and obtain a matching relationship between the preset fields and the feature identifiers.

9. The device according to claim 8, characterized in that The acquisition module is further used to acquire the parameter value of the target user identifier corresponding to the feature identifier; perform data standardization processing on the parameter value of the target user identifier corresponding to the feature identifier to obtain the feature value of the target user identifier corresponding to the feature identifier.

10. The device according to claim 9, characterized in that The sample user identifiers corresponding to the first category and the sample user identifiers corresponding to the second category are determined based on indicator parameters, and the indicator parameters include at least one of the account remaining value, the value transferred to the personal third-party account, the number of transfers to the personal third-party account, and the number of transfers to other people's accounts.

11. The device according to claim 9, characterized in that The feature selection conditions include: at least one of the information value being greater than a first threshold, the recall rate being greater than a second threshold, the statistical period being greater than a third threshold, and the business correlation coefficient being greater than a fourth threshold, and when the feature correlation coefficient of at least two of the candidate feature identifiers is greater than a fifth threshold, the candidate feature identifier with the highest recall rate is used as the target feature identifier.

12. The device according to claim 8, characterized in that The generation module is further used to generate a correlation coefficient matrix corresponding to the target user identity in each of the preset fields according to the characteristic information corresponding to the target user identity in each of the preset fields; obtain the characteristic roots and characteristic vectors of the correlation coefficient matrix corresponding to each of the preset fields; and generate a factor loading matrix corresponding to the target user identity in each of the preset fields according to the characteristic roots and characteristic vectors of the correlation coefficient matrix corresponding to each of the preset fields.

13. The device according to claim 8, characterized in that The determination module is further configured to determine the resource score of the target user identifier in each of the preset fields according to the feature information corresponding to the target user identifier in each of the preset fields and the feature coefficient of the feature information; The resource value of the target user identifier is determined according to the resource score of the target user identifier in each of the preset fields.

14. The device according to claim 13, characterized in that The determination module is further used to calculate the product between the characteristic information corresponding to the target user identifier in the preset field and the characteristic coefficient of the characteristic information; and add the products of the preset fields to obtain the resource score of the target user identifier in the preset field.

15. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

16. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images