Method and system for determining abnormal customer group based on burying point data and discriminant analysis

Abstract

The invention provides a method and system for determining an abnormal customer group based on burying point data and discriminant analysis, and relates to the technical field of abnormal customer group judgment, and the method comprises the steps: obtaining a target feature of a to-be-detected customer; inputting the target features into a constructed anomaly discrimination model to obtain two-dimensional coordinate points and a classification scatter diagram of the to-be-detected customer; according to the two-dimensional coordinate points and the classification scatter diagram, obtaining the category of the to-be-detected customer; the construction method of the anomaly discrimination model comprises the following steps: acquiring use behavior data of a customer; preprocessing the use behavior data to obtain customer feature data, vehicle feature data and micro-variable data; sorting the customer feature data, the vehicle feature data and the microvariable data to obtain wide table data; and based on a discriminant analysis method, drawing a scatter diagram and inputting the scatter diagram and the wide table data to obtain an anomaly discrimination model. The problems that in the prior art, loan field qualification information is prone to counterfeiting, the recognition rate is low, and white family crowds are difficult to evaluate are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of abnormal customer group determination, and in particular to a method and system for determining abnormal customer groups based on buried point data and discriminant analysis. Background Art

[0002] At present, with the economic downturn, fraudulent applications continue to appear in scenarios such as car installment / leasing, personal loans, and cash loans. Among them, packaged fraud and gang fraud have become the most hidden groups. Since the social security payments, provident fund payments, wages, and home purchases of these customers may be packaged and falsified, it is difficult to discover the risks using models based on this information. In addition, for white-household groups, due to the lack of historical data, it is impossible to build an application scoring card model based on historical data. Summary of the invention

[0003] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a method and system for determining abnormal customer groups based on buried data and discriminant analysis. The present invention solves the problems in the prior art that qualification information in the loan field is easy to falsify, the recognition rate is low, and it is difficult to evaluate the white household population.

[0004] To achieve the above object, the present invention provides the following solutions:

[0005] A method for determining abnormal customer groups based on buried data and discriminant analysis, comprising:

[0006] Obtain the target characteristics of the customers to be tested;

[0007] Input the target features into the constructed abnormality discrimination model to obtain the two-dimensional coordinate points and classification scatter plot of the customer to be tested;

[0008] Obtaining the category of the customer to be tested according to the two-dimensional coordinate points and the classification scatter plot;

[0009] The construction method of the abnormality discrimination model is:

[0010] Obtain customer usage behavior data;

[0011] Preprocessing the usage behavior data to obtain customer feature data, vehicle feature data and micro-variable data;

[0012] Arrange the customer characteristic data, vehicle characteristic data and micro-variable data to obtain wide table data;

[0013] Based on the discriminant analysis method, a scatter plot is drawn and the scatter plot and the wide table data are input to obtain an abnormality discrimination model.

[0014] Preferably, the target features include:

[0015] Applicant’s age, down payment ratio, applied vehicle price, total time spent by the end customer in filling out application materials, and length of time spent on a particular page.

[0016] Preferably, the microvariable data includes:

[0017] The total application time, page dwell time, page scrolling speed, number of terminal pop-ups, upload speed of various attachments, and shooting equipment information contained in the uploaded images.

[0018] Preferably, the abscissa calculation expression of the two-dimensional coordinate point is:

[0019] Function1 = -0.346*X1-0.525*X2+0.846*X3+0.613*X4+0.313*X5;

[0020] Among them, X1 is the applicant's age, X2 is the down payment ratio, X3 is the price of the applied vehicle, X4 is the total time it takes for the end customer to fill in the application materials, and X5 is the length of time it stays on a certain page.

[0021] Preferably, the vertical coordinate calculation expression of the two-dimensional coordinate point is:

[0022] Function2=0.039*X1+0.742*X2-0.386*X3+0.555*X4-0.216*X5.

[0023] A system for determining abnormal customer groups based on buried data and discriminant analysis, comprising:

[0024] A customer behavior data extraction module is used to obtain the behavior data of the customer to be tested;

[0025] A data processing module, used for preprocessing the behavior data to obtain customer characteristic data, vehicle characteristic data and micro-variable data;

[0026] A customer label processing module, used to organize the customer feature data, vehicle feature data and micro-variable data into wide table data and label each sample to obtain organized wide table data;

[0027] Model building module, used to build anomaly discrimination model based on sorted wide table data and discriminant analysis methods;

[0028] The conclusion processing module is used to input the target features of the customer to be tested into the abnormality discrimination model to obtain the category of the customer to be tested.

[0029] The present invention discloses the following technical effects:

[0030] The present invention provides a method and system for determining abnormal customer groups based on buried data and discriminant analysis, the method comprising:

[0031] Obtain the target features of the customer to be tested; input the target features into the constructed abnormal discrimination model to obtain the two-dimensional coordinate points and classification scatter plots of the customer to be tested; obtain the category of the customer to be tested according to the two-dimensional coordinate points and classification scatter plots; the construction method of the abnormal discrimination model is: obtain the customer's usage behavior data; pre-process the usage behavior data to obtain customer feature data, vehicle feature data and micro-variable data; sort the customer feature data, vehicle feature data and micro-variable data to obtain wide table data; based on the discriminant analysis method, draw a scatter plot and input the scatter plot and the wide table data to obtain an abnormal discrimination model. The present invention introduces these tiny variables of the customer's usage behavior data and uses the discriminant analysis method to construct an abnormal discrimination model, which improves the accuracy of judging the applicant's risk and provides a judgment basis for the judgment of the white household population. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0033] Figure 1 A flow chart of a method for determining abnormal customer groups based on buried point data and discriminant analysis provided by an embodiment of the present invention;

[0034] Figure 2 A schematic diagram of classification scatter points provided by an embodiment of the present invention;

[0035] Figure 3 A schematic diagram of display marks on a classification scatter diagram provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0038] like Figure 1 As shown, the present invention provides a method for determining abnormal customer groups based on buried point data and discriminant analysis, comprising:

[0039] Step 100: Obtain target features of the customer to be tested;

[0040] Step 200: input the target feature into the constructed abnormality discrimination model to obtain the two-dimensional coordinate points and classification scatter plot of the customer to be tested;

[0041] Step 300: Obtaining the category of the customer to be tested according to the two-dimensional coordinate points and the classification scatter plot;

[0042] The construction method of the abnormality discrimination model is:

[0043] Step 201: Obtaining customer usage behavior data;

[0044] Specifically, the customer's usage behavior data is extracted at the terminal, for example, (detailed information such as the model of the camera device, shooting time, location, and ISO speed and aperture setting used), etc., while the basic information of the device itself, location information, network WIFI, etc. are collected;

[0045] Step 202: pre-processing the usage behavior data to obtain customer feature data, vehicle feature data and micro-variable data;

[0046] Specifically, the data are extracted, sorted, cleaned, variables calculated and variables derived.

[0047] Data organization: Data merging: Merge data extracted from different sources (such as customer behavior data, device information, location information, etc.) to form a unified data set; Missing value processing: Check whether there are missing values ​​in the data set, and take corresponding measures according to the situation, such as filling missing values ​​(using mean, median or mode) or deleting samples with more missing values; Data format standardization: Ensure that the format of all data fields is consistent, such as date format, numerical type, etc., to facilitate subsequent analysis.

[0048] Data cleaning: Outlier detection: Identify and process outliers through statistical methods (such as Z-score, IQR), and decide whether to delete, correct or retain these outliers; Duplicate data processing: Check whether there are duplicate records in the data set, and choose to retain or delete duplicate data based on business needs; Data type conversion: Convert data fields to appropriate data types (such as converting strings to date types) to ensure data accuracy and availability.

[0049] Variable calculation: Derived variable creation: Calculate new variables based on existing data. For example: Calculate the frequency of use of customers (such as the number of photos taken per month). Calculate the average usage time of the device (such as the duration of taking photos). Categorical variable encoding: Convert categorical variables (such as device model, customer type) into numerical variables using methods such as One-Hot Encoding or Label Encoding.

[0050] Variable derivation: Time feature extraction: Extract useful time features from timestamps, such as the month, day of the week, hour, etc. of the photo, to analyze the time pattern of customer behavior. Aggregate feature calculation: Perform aggregate calculations on customer behavior data, such as calculating the average repayment amount of customers in the past three months. Calculate the incidence of fraudulent behavior of customers in a specific time period. Interaction feature creation: Create interaction features between variables, such as the interaction features between the frequency of customers using a specific device and their repayment performance, to analyze the impact of device use on repayment behavior.

[0051] Step 203: Arrange the customer characteristic data, vehicle characteristic data and micro-variable data to obtain wide table data;

[0052] Specifically, the data is organized into a wide table, and each sample is labeled with different categories, such as customer repayment performance (normal, overdue, bad, etc.), customer fraud type (normal, individual fraud, gang fraud, packaged fraud, etc.).

[0053] Step 204: Based on the discriminant analysis method, a scatter plot is drawn and the scatter plot and the wide table data are input to obtain an abnormality discrimination model.

[0054] Specifically, the sorted wide table data is input into the discriminant analysis method (discriminant analysis is to divide the known research objects into several types (or groups), and obtain the observation data of a batch of known samples of various types, on this basis, establish a discriminant formula according to certain criteria, and then discriminate and classify the samples of unknown types) to form a discriminant formula. Figure 2 shown.

[0055] Furthermore, the target features include:

[0056] Applicant’s age, down payment ratio, applied vehicle price, total time spent by the end customer in filling out application materials, and length of time spent on a particular page.

[0057] Furthermore, the microvariable data includes:

[0058] The total application time, page dwell time, page scrolling speed, number of terminal pop-ups, upload speed of various attachments, and shooting equipment information contained in the uploaded images.

[0059] Furthermore, the calculation expression of the horizontal coordinate of the two-dimensional coordinate point is:

[0060] Function1 = -0.346*X1-0.525*X2+0.846*X3+0.613*X4+0.313*X5;

[0061] Among them, X1 is the applicant's age, X2 is the down payment ratio, X3 is the price of the applied vehicle, X4 is the total time it takes for the end customer to fill in the application materials, and X5 is the length of time it stays on a certain page.

[0062] Furthermore, the vertical coordinate calculation expression of the two-dimensional coordinate point is:

[0063] Function2=0.039*X1+0.742*X2-0.386*X3+0.555*X4-0.216*X5.

[0064] This embodiment also provides a system for determining abnormal customer groups based on embedded point data and discriminant analysis, including:

[0065] A customer behavior data extraction module is used to obtain the behavior data of the customer to be tested;

[0066] A data processing module, used for preprocessing the behavior data to obtain customer characteristic data, vehicle characteristic data and micro-variable data;

[0067] A customer label processing module, used to organize the customer feature data, vehicle feature data and micro-variable data into wide table data and label each sample to obtain organized wide table data;

[0068] Model building module, used to build anomaly discrimination model based on sorted wide table data and discriminant analysis methods;

[0069] The conclusion processing module is used to input the target features of the customer to be tested into the abnormality discrimination model to obtain the category of the customer to be tested.

[0070] Furthermore, a new customer (Sample: 27 years old, 20%, 350,000, 3'21", 15") applies for the terminal, extracts the features required for discriminant analysis, inputs the abnormal discrimination model, and plots them on the scatter plot. Figure 3 As indicated by a red five-pointed star, calculate the Euclidean distance between the point of the new customer on the two-dimensional coordinate axis and the center point (center of gravity, etc.) of the three customer groups, and output the specific value. The customer group with the shortest distance to the new customer is determined to which category the applicant belongs.

[0071] Specifically, the present invention fully and effectively utilizes various types of information generated during the process of applicants applying for car loans or leasing vehicles, and through discriminant analysis methods, intuitively displays abnormal packaging customers and gang fraud customers, helping approval personnel to more intuitively observe which type of people the applicant is closer to, which can help approval personnel make approval decisions more directly; it can also directly output distance values ​​for automated judgment. At the same time, it is also applied to loan management and post-loan management, and judges the customer's repayment behavior based on the customer's activities at the terminal, and intervenes in a timely manner. At the same time, it can also be applied to channel management, and judges the channel's performance and operating conditions based on the basic characteristics of the channel and its activities at the terminal.

[0072] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0073] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for determining abnormal customer groups based on buried data and discriminant analysis, characterized in that: include: Obtain the target characteristics of the customers to be tested; Input the target features into the constructed abnormality discrimination model to obtain the two-dimensional coordinate points and classification scatter plot of the customer to be tested; Obtaining the category of the customer to be tested according to the two-dimensional coordinate points and the classification scatter plot; The construction method of the abnormality discrimination model is: Obtain customer usage behavior data; Preprocessing the usage behavior data to obtain customer characteristic data, vehicle characteristic data and micro-variable data; Arrange the customer characteristic data, vehicle characteristic data and micro-variable data to obtain wide table data; Based on the discriminant analysis method, a scatter plot is drawn and the scatter plot and the wide table data are input to obtain an abnormality discrimination model.

2. According to claim 1, a method for determining abnormal customer groups based on buried point data and discriminant analysis is characterized in that: The target features include: Applicant’s age, down payment ratio, applied vehicle price, total time spent by the end customer in filling out application materials, and length of time spent on a particular page.

3. The method for determining abnormal customer groups based on buried point data and discriminant analysis according to claim 1 is characterized in that: The microvariable data include: The total application time, page dwell time, page scrolling speed, number of terminal pop-ups, upload speed of various attachments, and shooting equipment information contained in the uploaded images.

4. The method for determining abnormal customer groups based on buried point data and discriminant analysis according to claim 1 is characterized in that: The calculation expression of the horizontal coordinate of the two-dimensional coordinate point is: Function1 = -0.346*X1-0.525*X2+0.846*X3+0.613*X4+0.313*X5; Among them, X1 is the applicant's age, X2 is the down payment ratio, X3 is the price of the applied vehicle, X4 is the total time it takes for the end customer to fill in the application materials, and X5 is the length of time it stays on a certain page.

5. The method for determining abnormal customer groups based on buried point data and discriminant analysis according to claim 4 is characterized in that: The calculation expression of the ordinate of the two-dimensional coordinate point is: Function2=0.039*X1+0.742*X2-0.386*X3+0.555*X4-0.216*X5.

6. A system for determining abnormal customer groups based on buried point data and discriminant analysis, characterized in that: include: A customer behavior data extraction module is used to obtain the behavior data of the customer to be tested; A data processing module, used for preprocessing the behavior data to obtain customer characteristic data, vehicle characteristic data and micro-variable data; A customer label processing module, used to organize the customer feature data, vehicle feature data and micro-variable data into wide table data and label each sample to obtain organized wide table data; Model building module, used to build anomaly discrimination model based on sorted wide table data and discriminant analysis methods; The conclusion processing module is used to input the target features of the customer to be tested into the abnormality discrimination model to obtain the category of the customer to be tested.

Images