User identification method, device, equipment and medium

Abstract

The present disclosure provides a user identification method, which can be applied to the fields of artificial intelligence and financial technology. The user identification method includes: obtaining asset information generated by a user to be identified within a first preset time interval and user attribute information of the user to be identified; filtering a target time window from the first preset time interval according to the asset information; obtaining target asset information features and target user attribute features associated with the initial user corresponding to the target time window; and inputting the target asset information features and target user attribute information corresponding to the target asset information features and target user attribute features into a target user identification model, and outputting the identification result of the target user. The present disclosure also provides a user identification device, equipment, storage medium and program product.

Description

Technical Field

[0001] The present disclosure relates to the fields of artificial intelligence and financial technology, and in particular to a user identification method, device, equipment, medium and program product. Background Art

[0002] With the continuous development of the Internet, transactions through the Internet have been widely used. In addition, with the deepening of digitalization, revitalizing existing users and improving marketing efficiency have become the key contents of Internet marketing. If effective product marketing can be carried out for existing users and the conversion of potential users to formal product users can be achieved, stable product user growth can be obtained. For example, in the domestic financial field, the long tail theory also has a more beneficial reference significance and far-reaching impact on the development of retail business of banks.

[0003] In the process of implementing the embodiments of the present disclosure, it is found that potential users are easily overlooked according to the long tail theory, and the aggregation of these potential users will also affect the marketing efficiency of Internet marketing. Summary of the invention

[0004] In view of the above problems, the present disclosure provides a user identification method, apparatus, device, medium and program product.

[0005] According to a first aspect of the present disclosure, there is provided a user identification method, comprising:

[0006] Acquire asset information generated by the user to be identified within a first preset time interval and user attribute information of the user to be identified;

[0007] Filtering a target time window from a first preset time interval according to the asset information;

[0008] Acquire target asset information features and target user attribute features associated with the initial user corresponding to the target time window; and

[0009] The target asset information and target user attribute information corresponding to the target asset information characteristics and the target user attribute characteristics are input into the target user identification model, and the identification result of the target user is output.

[0010] According to an embodiment of the present disclosure, filtering a target time window from a first preset time interval according to the asset information includes:

[0011] Determine n time windows from the first preset time interval, where n is a positive integer greater than or equal to 2;

[0012] Determine a preset observation window and a preset performance window corresponding to the i-th time window, wherein i≤n, the preset observation window is calculated based on the i-th time window and the first interval duration; the preset performance window is calculated based on the preset observation window and the second interval duration;

[0013] When the asset information of the corresponding user in the preset observation window and the preset performance window meets the asset threshold, the i-th time window is taken as the target time window.

[0014] According to an embodiment of the present disclosure, the target user recognition model is obtained by pre-training, and the pre-training method includes:

[0015] Acquire sample user asset information and sample user attribute information of the sample user generated by the sample user within a second preset time interval;

[0016] Filtering a target sample time window from a second preset time interval according to the sample user asset information;

[0017] Obtaining target sample user asset information features and target sample user attribute features associated with the initial sample user corresponding to the target sample time window;

[0018] Determine a training set according to the target sample user asset information characteristics and the target sample user attribute information corresponding to the target sample user attribute characteristics;

[0019] Input the training set into a predetermined initial recognition model and output the initial recognition result;

[0020] Based on the initial recognition results and the actual results, the parameters of the initial recognition model are adjusted.

[0021] According to an embodiment of the present disclosure, the predetermined initial recognition model is determined according to a preset quantitative index;

[0022] Among them, the preset quantitative indicators include at least: indicators for evaluating the model's discrimination and indicators for evaluating the model's ranking ability.

[0023] According to an embodiment of the present disclosure, the target asset information characteristics include at least one of the following: characteristics characterizing financial management, characteristics characterizing transactions, characteristics characterizing credit, and characteristics characterizing income.

[0024] According to an embodiment of the present disclosure, before acquiring the asset information generated by the user to be identified within the first preset time interval and the user attribute information of the user to be identified, the method further includes:

[0025] Determine the type of user to be identified based on the asset information;

[0026] When the type of the user to be identified satisfies the long-tail customer, the asset information generated by the user to be identified within a first preset time interval and the user attribute information of the user to be identified are obtained.

[0027] A second aspect of the present disclosure provides a user identification device, comprising:

[0028] A first acquisition module, used to acquire asset information generated by the user to be identified within a first preset time interval and user attribute information of the user to be identified;

[0029] A screening module, used for screening a target time window from a first preset time interval according to the asset information;

[0030] A second acquisition module is used to acquire target asset information characteristics and target user attribute characteristics associated with the initial user corresponding to the target time window; and

[0031] The identification module is used to input the target asset information characteristics and the target user attribute information corresponding to the target asset information characteristics and the target user attribute characteristics into the target user identification model, and output the identification result of the target user.

[0032] A third aspect of the present disclosure provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the above-mentioned user identification method.

[0033] The fourth aspect of the present disclosure further provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the above-mentioned user identification method.

[0034] The fifth aspect of the present disclosure also provides a computer program product, including a computer program, which implements the above-mentioned user identification method when executed by a processor.

[0035] According to the embodiment of the present disclosure, based on the asset information generated by the user to be identified obtained within the first preset time interval, the target time window is screened from the time window of the first preset time interval; in addition, by obtaining the target features under the target time window, the accuracy of the target user identification model in distinguishing the target user is improved. Products and services can be designed for target users, customer stickiness can be enhanced, and marketing efficiency can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The above contents and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0037] Figure 1A diagram schematically illustrates an application scenario of a user identification method, apparatus, device, medium, and program product according to an embodiment of the present disclosure;

[0038] Figure 2 A flowchart of a user identification method according to an embodiment of the present disclosure is schematically shown;

[0039] Figure 3 A flowchart of a method for selecting a target time window from a first preset time interval according to asset information according to an embodiment of the present disclosure is schematically shown;

[0040] Figure 4 Schematically showing a target user identification model obtained by pre-training according to an embodiment of the present disclosure, a flow chart of a pre-training method;

[0041] Figure 5 A structural block diagram of a user identification device according to an embodiment of the present disclosure is schematically shown; and

[0042] Figure 6 A block diagram of an electronic device suitable for implementing a user identification method according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0043] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0044] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0045] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0046] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0047] In the technical solution disclosed in the present invention, the acquisition, storage and application of user personal information involved are in compliance with the provisions of relevant laws and regulations, necessary confidentiality measures are taken, and do not violate public order and good morals.

[0048] In the technical solution of the embodiment of the present disclosure, the user's authorization or consent is obtained before obtaining or collecting the user's personal information.

[0049] In the process of implementing the embodiments of the present disclosure, it is found that in terms of user positioning, commercial banks are no different from other traditional commercial enterprises. They are used to classifying users into high and low levels and maintaining the inertial thinking of the "80 / 20 rule". There are also potential users among the ignored long-tail users. Gathering these potential users may also be a huge market. It is possible that this "long neglected tail" of long-tail users will also support the future development of domestic commercial banks to a large extent.

[0050] The business strategies of traditional commercial banks often believe in the 80 / 20 rule, that is, 20% of the people own 80% of the social wealth. This common choice is often caused by the imbalance of the input-output ratio, which leads to the pattern tendency of the user operation process. The tendency to invest in high-quality users in turn reacts to the existence of this imbalance, and gradually long-tail users have become the weak area of ​​traditional commercial bank services. With the development of financial technology and the popularization of digital products, the realization of the user-centric concept has become possible. Especially for long-tail users, with the support of big data, how to dig out pseudo-long-tail users, that is, potential users, from traditional long-tail users can not only share lower operating costs, but also greatly improve marketing efficiency, help increase revenue, and provide users with efficient, comfortable and customized service experience.

[0051] An embodiment of the present disclosure provides a user identification method, including: obtaining asset information generated by a user to be identified within a first preset time interval and user attribute information of the user to be identified; filtering a target time window from the first preset time interval based on the asset information; obtaining target asset information features and target user attribute features associated with the initial user corresponding to the target time window; and inputting the target asset information features and target user attribute information corresponding to the target asset information features and the target user attribute features into a target user identification model, and outputting an identification result of the target user.

[0052] Figure 1 The application scenario diagram of the user identification method, apparatus, device, medium and program product according to the embodiments of the present disclosure is schematically shown.

[0053] like Figure 1 As shown, the application scenario 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0054] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only examples).

[0055] The terminal devices 101 , 102 , and 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0056] The server 105 may be a server that provides various services, such as a background management server (only an example) that provides support for websites browsed by users using the terminal devices 101, 102, and 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0057] It should be noted that the user identification method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the user identification device provided in the embodiment of the present disclosure can generally be set in the server 105. The user identification method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the user identification device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105.

[0058] The user identification method provided in the embodiment of the present disclosure may also be executed by the terminal devices 101, 102, 103. Accordingly, the user identification device provided in the embodiment of the present disclosure may also be generally arranged in the terminal devices 101, 102, 103. The user identification method provided in the embodiment of the present disclosure may also be executed by other terminals different from the terminal devices 101, 102, 103. Accordingly, the user identification device provided in the embodiment of the present disclosure may also be arranged in other terminals different from the terminal devices 101, 102, 103.

[0059] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0060] The following will be based on Figure 1 The scene described by Figure 2 to Figure 4 The user identification method of the disclosed embodiment is described in detail.

[0061] Figure 2 The flowchart of the user identification method according to the embodiment of the present disclosure is schematically shown.

[0062] like Figure 2 As shown, the user identification method 200 of this embodiment includes operations S201 to S204.

[0063] In operation S201, asset information generated by a user to be identified within a first preset time interval and user attribute information of the user to be identified are obtained.

[0064] According to an embodiment of the present disclosure, the authorization of the user to be identified for the asset information and the user attribute information can be obtained; after obtaining the authorization, the asset information and the user attribute information are obtained. The asset information may include, but is not limited to: physical precious metal information, financial information, consumption information, salary information, credit card information, loan information, bank card information, real estate information, other physical information, etc. The user attribute information may include, but is not limited to: basic user information, user information associated with the user. The first preset time interval may include a period of time in the past, for example, may include but is not limited to: the past 12 months, the past 15 months, the past 18 months, etc. The type of the user to be identified may be the long-tail customer referred to in the traditional financial field.

[0065] In operation S202, a target time window is filtered from a first preset time interval according to the asset information.

[0066] According to an embodiment of the present disclosure, multiple time windows can be obtained based on a first preset time interval; a preset observation window and a preset performance window corresponding to each time window are determined respectively; wherein the preset observation window can be used to determine whether the average asset information belonging to the preset observation window meets the asset threshold. The preset performance window can be used to determine whether the average asset information belonging to the preset performance window meets the asset threshold. If the asset information of the user corresponding to the preset observation window and the preset performance window of a time window both meet the asset threshold, the time window can be used as the target time window. It should be noted that the target time window here may include multiple time windows. Among them, the time window can be a time period with a length of months.

[0067] The preset observation window is calculated based on the time window and the first interval duration, and the preset performance window is calculated based on the preset observation window and the second interval duration. Both the first interval duration and the second interval duration can be several months.

[0068] For example, the first preset time interval is from January 1, 2021 to November 30, 2021. Three time windows can be determined from this time period, each time window is one month long, such as 202101, 202102, and 202103. The time window of 202101 can correspond to the preset observation window from January 1, 2021 to June 30, 2021 and the preset performance window from July 1, 2021 to September 30, 2021. The time window of 202102 can correspond to the preset observation window from February 1, 2021 to July 31, 2021 and the preset performance window from August 1, 2021 to October 31, 2021. The time window of 202103 may correspond to the preset observation window from March 1, 2021 to August 31, 2021 and the preset performance window from September 1, 2021 to November 30, 2021. It should be noted that the specific time in this embodiment is only an example and is not intended to limit the present disclosure.

[0069] In operation S203, target asset information characteristics and target user attribute characteristics associated with the initial user corresponding to the target time window are obtained.

[0070] According to the embodiment of the present disclosure, for a target time window, target asset information features and target user attribute features associated with an initial user belonging to the target time window may be acquired.

[0071] According to an embodiment of the present disclosure, the target asset information characteristics include at least one of the following: characteristics characterizing financial management, characteristics characterizing transactions, characteristics characterizing credit, and characteristics characterizing income.

[0072] According to an embodiment of the present disclosure, the target user attribute features include address association features characterizing the target user and users associated with the target user, and personal information association features characterizing the target user and users associated with the target user.

[0073] In operation S204, the target asset information and the target user attribute information corresponding to the target asset information characteristics and the target user attribute characteristics are input into a target user identification model, and an identification result of the target user is output.

[0074] According to the embodiment of the present disclosure, the target asset information and the target user attribute information are data under the target asset information feature and the target user attribute feature dimension respectively. The target asset information and the target user attribute information can be input into the target user identification model, and the identification result of the target user can be output. Among them, the identification result of the target user can be the target user or the non-target user. The target user identification model can be obtained by pre-training.

[0075] According to the disclosed embodiment, based on the asset information generated by the user to be identified obtained within the first preset time interval, the target time window is screened from the time window of the first preset time interval; in addition, the accuracy of the target user identification model in distinguishing the target user is improved by obtaining the target features under the target time window. Products and services can be designed for target users, customer stickiness can be enhanced, and marketing efficiency can be improved.

[0076] According to an embodiment of the present disclosure, before operation S204, the process may further include: excluding initial users whose household age does not meet the preset household age according to the behavioral characteristics of the initial users. For example, initial users whose household age does not meet the preset household age (e.g., 6 months) may be excluded for credit cards or bank cards.

[0077] It may also include: directly excluding initial users whose monthly average daily asset balance is 0 within a preset observation window.

[0078] According to an embodiment of the present disclosure, before acquiring the asset information generated by the user to be identified within the first preset time interval and the user attribute information of the user to be identified, the method further includes:

[0079] Determine the type of user to be identified based on the asset information;

[0080] When the type of the user to be identified satisfies the long-tail customer, the asset information generated by the user to be identified within a first preset time interval and the user attribute information of the user to be identified are obtained.

[0081] Figure 3 A flowchart of a method for screening a target time window from a first preset time interval based on asset information according to an embodiment of the present disclosure is schematically shown.

[0082] like Figure 3 As shown, the method 300 of selecting a target time window from a first preset time interval according to asset information in this embodiment includes operations S301 to S303.

[0083] In operation S301, n time windows are determined from a first preset time interval, where n is a positive integer greater than or equal to 2.

[0084] According to an embodiment of the present disclosure, the time window may be divided according to the length of the first preset time interval. For example, if the first preset time interval is 1 year, the time window may be a time period in months.

[0085] In operation S302, a preset observation window and a preset performance window corresponding to the i-th time window are determined, where i≤n, the preset observation window is calculated based on the i-th time window and the first interval duration; the preset performance window is calculated based on the preset observation window and the second interval duration.

[0086] According to an embodiment of the present disclosure, if the i-th time window is a month of a certain year, the preset observation window may be a start time of the month and end after a first interval. The preset performance window may be a second interval after the end time of the preset observation window.

[0087] For example, the i-th time window is July 2020, which can be expressed as 202007. The preset observation window can be from the beginning of July 2020 to the end of November 2020; the preset performance window can be from the beginning of January 2021 to the end of March 2021.

[0088] In operation S303, when the asset information of the corresponding user in the preset observation window and the preset performance window both meet the asset threshold, the i-th time window is used as the target time window.

[0089] According to an embodiment of the present disclosure, the asset threshold can be determined based on the accuracy of the target user identification model actually selected. For example, the asset threshold can be 10,000. The preset observation window can be further divided into multiple sub-observation windows in units of months, and the monthly average assets of each sub-observation window can be compared with the asset threshold. When the monthly average assets of at most two sub-observation windows meet the asset threshold, it can be determined that the asset information of the corresponding user in the preset observation window meets the asset threshold. The preset performance window can be further divided into multiple sub-performance windows in units of months, and the monthly average assets of each sub-performance window can be compared with the asset threshold. When the monthly average assets of each sub-performance window meet the asset threshold, it can be determined that the asset information of the corresponding user in the preset performance window meets the asset threshold.

[0090] According to the embodiments of the present disclosure, by presetting an observation window and a performance window and screening a target time window from multiple time windows, the stability and purity of target users identified later can be ensured, and the phenomenon of low concentration of target users in the selected window can be effectively avoided.

[0091] Figure 4 The flowchart of the pre-training method is schematically shown in which a target user recognition model according to an embodiment of the present disclosure is obtained through pre-training.

[0092] like Figure 4 As shown, the target user recognition model of this embodiment is obtained through pre-training, and the pre-training method 400 includes operations S401 to S406.

[0093] In operation S401, sample user asset information and sample user attribute information of the sample user generated within a second preset time interval are obtained.

[0094] According to an embodiment of the present disclosure, the second preset time interval may be a period of time in the past, for example, the past year, the past 10 months, etc. The sample user asset information generated by all sample users from September 1, 2020 to September 30, 2021 and the sample user attribute information of the sample users may be obtained.

[0095] In operation S402, a target sample time window is filtered from a second preset time interval according to the sample user asset information.

[0096] According to an embodiment of the present disclosure, n sample time windows may be divided according to the length of the second preset time interval in the above operation S401. Where n is a positive integer greater than or equal to 2. The sample preset observation window and the sample preset performance window corresponding to the i-th sample time window are determined. When the asset information of the corresponding sample user in the sample preset observation window and the sample preset performance window both meet the asset threshold, the i-th sample time window is used as the target sample time window.

[0097] For example, the asset threshold can be determined based on the efficiency of mining pseudo long-tail customer groups from different asset levels through statistical analysis of long-tail customer groups. As shown in the statistical analysis results in Table 1 below, with June 2020 as the observation point, among customers with a monthly average daily asset level of more than 10,000 in the past six months, the monthly average daily asset level has declined to varying degrees in the following six months. Moreover, customers at the (50,000, 100,000) and (100,000, 200,000) levels have been downgraded to a large extent, and the pattern of migration to different levels is relatively vague.

[0098] Table 1

[0099]

[0100] In summary, the asset threshold can be determined as 10,000 (i.e. 1w).

[0101] Table 2

[0102]

[0103] The sample time windows can be determined as 202009, 202010, 202011, and 202012, as shown in Table 2 above. The sample preset observation window can be further divided into 6 sub-observation windows in units of months. The monthly average daily assets of each sub-sample observation window can be compared with the asset threshold. If the monthly average daily assets of at most two sub-sample observation windows meet the asset threshold, it can be determined that the asset information of the corresponding user in the sample preset observation window meets the asset threshold. The sample preset performance window can be further divided into 3 sub-sample performance windows in units of months. The monthly average daily assets of each sub-sample performance window can be compared with the asset threshold. If the monthly average daily assets of each sub-sample performance window meet the asset threshold, it can be determined that the asset information of the corresponding user in the sample preset performance window meets the asset threshold. The three sample time windows of 202010, 202011, and 202012 are obtained as shown in Table 2.

[0104] It should be noted that if the sample preset observation window is further divided into 6 sub-observation windows with monthly units, and the monthly average daily assets of each sub-sample observation window are compared with the asset threshold, if the monthly average daily assets of three sub-sample observation windows meet the asset threshold, it is determined that the asset information of the corresponding user in the sample preset observation window meets the asset threshold; if the sample preset performance window is further divided into 3 sub-sample performance windows with monthly units, and the monthly average daily assets of each sub-sample performance window are compared with the asset threshold, if the monthly average daily assets of each sub-sample performance window meet the asset threshold, it is determined that the asset information of the corresponding user in the sample preset performance window meets the asset threshold. In this case, it can be found through analysis of the obtained sample user asset information that the concentration of target users under this condition is low, which is not conducive to improving the marketing efficiency of target users after identifying target users in the later stage.

[0105] In operation S403, the target sample user asset information characteristics and the target sample user attribute characteristics associated with the initial sample user corresponding to the target sample time window are obtained.

[0106] According to an embodiment of the present disclosure, for a target sample time window, target sample asset information features and target sample user attribute features associated with a target sample user belonging to the target sample time window may be acquired.

[0107] According to an embodiment of the present disclosure, the target sample asset information characteristics may include at least one of the following: characteristics characterizing the sample financial management class, characteristics characterizing the sample transaction class, characteristics characterizing the sample credit class, and characteristics characterizing the sample income class.

[0108] According to an embodiment of the present disclosure, the target sample user attribute features include address association features characterizing the target sample user and sample users associated with the target sample user, and personal information association features characterizing the target sample user and sample users associated with the target sample user.

[0109] According to an embodiment of the present disclosure, in addition to the above operation S403, the initial sample users whose household age does not meet the preset household age can be excluded according to the behavioral performance characteristics of the initial sample users. For example, for credit cards or bank cards, the initial sample users whose household age does not meet the preset household age (for example, 6 months) can be excluded.

[0110] According to an embodiment of the present disclosure, in addition to the above operation S403, initial sample users whose monthly average daily asset balance is 0 within the sample preset observation window may also be excluded.

[0111] In operation S404, a training set is determined according to the target sample user asset information characteristics and the target sample user attribute information corresponding to the target sample user attribute characteristics.

[0112] According to the embodiment of the present disclosure, the target sample user asset information and the target sample user attribute information are respectively the target sample user asset information features and the sample data under the target sample user attribute feature dimensions.

[0113] Three types of sample sets can be selected from the sample data: training set (Train), validation set (V) and test set (T). The training set is the actual data set for model establishment. The validation set is used to check the robustness of the model established using the training set. The model should be robust enough to have the same predictive ability on homogeneous data from different sources. The test set is used to perform out-of-sample verification of the learned model in order to verify the generalization ability of the model. When modeling, the stratified random sampling sampling method is used to prepare the modeling samples. The sampling method of stratified sampling is different from random sampling. Stratified sampling first determines the sample category according to the needs of the model, determines the number of samples for each category, and then randomly selects the required samples within each category. The sampling ratio of each category is different, and the corresponding sample weight is set. The stratified sampling method can ensure that the number of samples of each category in the modeling sample is sufficient.

[0114] For example, we can first extract a sample of all target sample users or a sample of 400,000 target sample users; then randomly extract a sample of non-target sample users that is 4 times the number of the above samples. Finally, the stratified sampling samples are split into training set, validation set, and test set in a ratio of 7:2:1.

[0115] In operation S405, the training set is input into a predetermined initial recognition model, and an initial recognition result is output.

[0116] According to an embodiment of the present disclosure, the predetermined initial recognition model is determined according to preset quantitative indicators; wherein the preset quantitative indicators include at least: an indicator for evaluating the model discrimination and an indicator for evaluating the model ranking ability.

[0117] According to the embodiments of the present disclosure, the algorithm models of LightGBM, XGBoost, GBM and GLM can be selected as alternative models. In the process of model comparison, the indicators used to evaluate the model discrimination (such as KS) and the indicators used to evaluate the model ranking ability (such as AUC) can be used as the main quantitative indicators. At the same time, the ROC graph and KS graph can also be referred to, with the precision rate as a secondary quantitative indicator. The predetermined initial recognition model is obtained by comprehensively considering the results of the above comparison criteria of the training set, validation set and test set.

[0118] For example, in the recognition model of target users below 10,000, the samples with the target sample time window of 202012 can be reserved in advance as inter-period samples (OOT) for further verification of the generalization ability of the model to be produced. The algorithm models of LightGBM, XGBoost, GBM and GLM are respectively used as candidate models for the recognition model of target users below 10,000. The performance results are shown in Table 3 below. It can be seen that the KS of the GBM algorithm model is as high as 0.51 and the AUC is as high as 0.84. The model has the best effect, so the GBM algorithm model can be selected as the predetermined initial recognition model.

[0119] Among them, VKS represents the KS indicator for the validation set; TKS represents the KS indicator for the test set, and OOTKS represents the KS indicator for inter-period samples. VAUC represents the AUC indicator for the validation set; TAUC represents the AUC indicator for the test set, and OOTAUC represents the AUC indicator for inter-period samples.

[0120] Table 3

[0121] Target user identification model algorithm VKS TKS OOTKS VAUC TAUC OOTAU 1W or less GBM 0.51 0.51 0.47 0.84 0.84 0.82 1W or less XGBoost 0.43 0.43 0.38 0.78 0.78 0.76 1W or less LGB 0.37 0.34 0.32 0.75 0.73 0.72 1W or less GLM 0.44 0.43 0.39 0.79 0.78 0.76

[0122] In operation S406, based on the initial recognition result and the true result, the parameters of the initial recognition model are adjusted.

[0123] According to the embodiment of the present disclosure, the target sample time window is screened from the sample time window of the second preset time interval according to the asset information generated by the sample user obtained in the second preset time interval; in addition, the training set, the validation set and the test set are obtained by obtaining the target sample features under the target sample time window. The accuracy of the initial recognition model in distinguishing the target sample users is increased by predetermining the initial recognition model, and the target user recognition model finally obtained can accurately identify the target user, so as to design products and provide services for the target user, enhance customer stickiness and improve marketing efficiency.

[0124] Based on the above user identification method, the present disclosure also provides a user identification device. Figure 5 The device is described in detail.

[0125] Figure 5 The structure block diagram of a user identification device according to an embodiment of the present disclosure is schematically shown.

[0126] like Figure 5 As shown, the user identification device 500 of this embodiment includes a first acquisition module 510 , a screening module 520 , a second acquisition module 530 and an identification module 540 .

[0127] The first acquisition module 510 is used to acquire the asset information generated by the user to be identified and the user attribute information of the user to be identified within a first preset time interval. In one embodiment, the first acquisition module 510 can be used to perform the operation S201 described above, which will not be repeated here.

[0128] The screening module 520 is used to screen the target time window from the first preset time interval according to the asset information. In one embodiment, the screening module 520 can be used to perform the operation S202 described above, which will not be described in detail here.

[0129] The second acquisition module 530 is used to acquire the target asset information characteristics and target user attribute characteristics associated with the initial user corresponding to the target time window. In one embodiment, the second acquisition module 530 can be used to perform the operation S203 described above, which will not be repeated here.

[0130] The identification module 540 is used to input the target asset information characteristics and the target user attribute information corresponding to the target asset information characteristics and the target user attribute characteristics into the target user identification model, and output the identification result of the target user. In one embodiment, the identification module 540 can be used to perform the operation S204 described above, which will not be repeated here.

[0131] According to an embodiment of the present disclosure, the user identification device 500 may further include a first exclusion module and a second exclusion module.

[0132] The first exclusion module is used to exclude initial users whose household age does not meet a preset household age according to the behavioral performance characteristics of the initial users.

[0133] The second exclusion module is used to directly exclude initial users whose monthly average daily asset balance is 0 within a preset observation window.

[0134] According to an embodiment of the present disclosure, the user identification device 500 may further include a type determination module.

[0135] The type determination module is used to determine the type of user to be identified based on asset information.

[0136] According to an embodiment of the present disclosure, any multiple modules of the first acquisition module 510, the screening module 520, the second acquisition module 530 and the identification module 540 can be combined in one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present disclosure, at least one of the first acquisition module 510, the screening module 520, the second acquisition module 530 and the identification module 540 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware and firmware or in a suitable combination of any of them. Alternatively, at least one of the first acquisition module 510 , the screening module 520 , the second acquisition module 530 , and the identification module 540 may be at least partially implemented as a computer program module, and when the computer program module is executed, a corresponding function may be performed.

[0137] Figure 6 A block diagram of an electronic device suitable for implementing a user identification method according to an embodiment of the present disclosure is schematically shown.

[0138] like Figure 6As shown, the electronic device 600 according to an embodiment of the present disclosure includes a processor 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage part 608 into a random access memory (RAM) 603. The processor 601 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 601 may also include an onboard memory for caching purposes. The processor 601 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0139] In RAM 603, various programs and data required for the operation of electronic device 600 are stored. Processor 601, ROM 602 and RAM 603 are connected to each other via bus 604. Processor 601 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM 602 and / or RAM 603. It should be noted that the program can also be stored in one or more memories other than ROM 602 and RAM 603. Processor 601 can also perform various operations of the method flow according to the embodiment of the present disclosure by executing the program stored in the one or more memories.

[0140] According to an embodiment of the present disclosure, the electronic device 600 may further include an input / output (I / O) interface 605, which is also connected to the bus 604. The electronic device 600 may further include one or more of the following components connected to the I / O interface 605: an input portion 606 including a keyboard, a mouse, etc.; an output portion 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 608 including a hard disk, etc.; and a communication portion 609 including a network interface card such as a LAN card, a modem, etc. The communication portion 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed, so that a computer program read therefrom is installed into the storage portion 608 as needed.

[0141] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.

[0142] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device. For example, according to an embodiment of the present disclosure, a computer-readable storage medium may include the ROM 602 and / or RAM 603 described above and / or one or more memories other than ROM 602 and RAM 603.

[0143] The embodiment of the present disclosure also includes a computer program product, which includes a computer program, and the computer program contains program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the method provided by the embodiment of the present disclosure.

[0144] The above functions defined in the system / device of the embodiment of the present disclosure are performed when the computer program is executed by the processor 601. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0145] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 609, and / or installed from a removable medium 611. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0146] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 609, and / or installed from the removable medium 611. When the computer program is executed by the processor 601, the above functions defined in the system of the embodiment of the present disclosure are performed. According to the embodiment of the present disclosure, the system, device, apparatus, module, unit, etc. described above can be implemented by a computer program module.

[0147] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0148] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0149] It will be appreciated by those skilled in the art that the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways, even if such combinations and / or combinations are not explicitly described in the present disclosure. In particular, the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways without departing from the spirit and teachings of the present disclosure. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0150] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above separately, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. The scope of the present disclosure is defined by the attached claims and their equivalents. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A user identification method, comprising: Acquire asset information generated by a user to be identified within a first preset time interval and user attribute information of the user to be identified; Filtering a target time window from the first preset time interval according to the asset information; Acquire target asset information characteristics and target user attribute characteristics associated with the initial user corresponding to the target time window; and Input the target asset information characteristics and the target user attribute information corresponding to the target asset information characteristics and the target user attribute characteristics into a target user identification model, and output an identification result of the target user; The step of selecting a target time window from the first preset time interval according to the asset information includes: Determine n time windows from the first preset time interval, where n is a positive integer greater than or equal to 2; Determine a preset observation window and a preset performance window corresponding to the i-th time window, wherein i≤n, the preset observation window is calculated based on the i-th time window and the first interval duration; the preset performance window is calculated based on the preset observation window and the second interval duration; When the asset information of the corresponding user in the preset observation window and the preset performance window both meet the asset threshold, the i-th time window is used as the target time window.

2. The method according to claim 1, wherein: The target user identification model is obtained by pre-training, and the pre-training method includes: Acquire sample user asset information generated by the sample user within a second preset time interval and sample user attribute information of the sample user; Filtering a target sample time window from the second preset time interval according to the sample user asset information; Acquire target sample user asset information characteristics and target sample user attribute characteristics associated with the initial sample user corresponding to the target sample time window; Determine a training set according to the target sample user asset information feature and the target sample user asset information and the target sample user attribute information corresponding to the target sample user attribute feature; Inputting the training set into a predetermined initial recognition model and outputting an initial recognition result; Based on the initial recognition result and the true result, the parameters of the initial recognition model are adjusted.

3. The method according to claim 2, wherein: The predetermined initial recognition model is determined according to a preset quantitative index; The preset quantitative indicators include at least: an indicator for evaluating the model discrimination and an indicator for evaluating the model ranking ability.

4. The method according to claim 1, wherein: The target asset information characteristics include at least one of the following: characteristics representing financial management, characteristics representing transactions, characteristics representing credit, and characteristics representing income.

5. The method according to claim 1, before acquiring the asset information generated by the user to be identified within the first preset time interval and the user attribute information of the user to be identified, further comprises: Determining the type of the user to be identified according to the asset information; In the case that the type of the user to be identified satisfies the long-tail customer, the asset information generated by the user to be identified within a first preset time interval and the user attribute information of the user to be identified are obtained.

6. A user identification device, comprising: A first acquisition module, used to acquire asset information generated by a user to be identified within a first preset time interval and user attribute information of the user to be identified; A screening module, configured to screen a target time window from the first preset time interval according to the asset information; A second acquisition module is used to acquire target asset information characteristics and target user attribute characteristics associated with the initial user corresponding to the target time window; and An identification module, used to input the target asset information characteristics and the target user attribute information corresponding to the target asset information characteristics and the target user attribute characteristics into a target user identification model, and output an identification result of the target user; The step of selecting a target time window from the first preset time interval according to the asset information includes: Determine n time windows from the first preset time interval, where n is a positive integer greater than or equal to 2; Determine a preset observation window and a preset performance window corresponding to the i-th time window, wherein i≤n, the preset observation window is calculated based on the i-th time window and the first interval duration; the preset performance window is calculated based on the preset observation window and the second interval duration; When the asset information of the corresponding user in the preset observation window and the preset performance window both meet the asset threshold, the i-th time window is used as the target time window.

7. An electronic device comprising: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors execute the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the method according to any one of claims 1 to 5.

9. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images