CS-PNN-based customer credit risk assessment method and system

A customer and credit technology, applied in the risk control field of the Internet financial industry, can solve the problems of long learning process, easy to fall into local optimum, poor robustness, etc.

CN112529683APending Publication Date: 2021-03-19百维金科(上海)信息科技有限公司
9 Cites 3 Cited by

Patent Information

Authority / Receiving Office
CN · China
Current Assignee / Owner
Publication Date
2021-03-19

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention relates to the technical field of risk control of the Internet financial industry, in particular to a CS-PNN-based customer credit risk assessment method and system. Compared with BP andRBF neural networks, the PNN fuses a Bayesian decision theory and density function estimation on the basis of a radial basis function, the method has the advantages of simple network structure, few adjustment parameters, short operation time, no local minimum point and the like; compared with GA, PSO, ACO and other optimization algorithms, the CS algorithm searches for a global optimal solution by simulating the combination of the parasitic propagation behavior of the valley bird nest and the Levy flight search principle, has the advantages of being few in parameter setting, high in convergence speed, high in universality and robustness, easy to implement and the like, and can efficiently balance local search and global search of the algorithm; the CSPNN model obtained by optimizing the smoothing factor of the PNN by the CS has the advantages of simple network structure, high convergence rate, good fault tolerance, high robustness, high classification accuracy, strong sample appendingcapability and the like, and can meet the requirement of real-time credit risk assessment of a loan system.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention belongs to the technical field of risk control in the Internet financial industry, and specifically provides a CS-PNN-based customer credit risk assessment method and system. Background technique

[0002] With the rapid development of Internet finance, the assessment of customer credit risk has become an important research field. Credit risk assessment uses information submitted by loan or credit card applicants and information provided by third parties to calculate the applicant's credit risk, classify them into different risk levels, and use this as the basis for loan or credit card approval.

[0003] Credit evaluation is essentially a classification problem in pattern recognition. According to the characteristics of applicants, such as age, gender, marital status and income, etc., machine learning methods are used to classify applicants into customers with different credit levels. When training the model, discover the rules based on t...

Examples

Embodiment 1

[0075] see figure 1 , the present invention provides a technical solution:

[0076] A method and system for assessing customer credit risk based on CS-PNN, comprising the following steps:

[0077] S1. Construct sample data, sample customers with existing loan performance as modeling samples, and collect customer credit characteristic data;

[0078] S2. Perform data preprocessing on the collected data, use the Min-Max method to normalize the preprocessed data, and divide it into a training set and a test set according to a ratio of 7:3;

[0079] S3. In the training set, first select 10 feature vectors that can most affect the repayment status by logistic regression or random forest as input, and use whether the repayment is overdue as output to establish a PNN prediction model;

[0080] S4. Use the CS algorithm to optimize the smoothing factor of PNN. The optimization algorithm aims at the accuracy of the model, obtains the best smoothing factor through iterative optimization...

Embodiment 2

[0143] The same parts of Embodiment 2 and Embodiment 1 will not be described in detail. The difference is: in S1, the customers with existing loan performance are sampled as modeling samples, and the credit characteristic data of customers are collected. The credit characteristic data includes personal basic information, Operational behavior data and third-party data, this setting is conducive to the collection of user data,

[0144] In S2, in the input of the neural network, due to the different units of each input, the order of magnitude differs greatly; if the direct input is used, it is easy to make the neuron training saturated, so before the input training, the data must be Normalize to make it at the same quantitative level, and use the Min-Max method to normalize the preprocessed data. The calculation formula is as follows:

[0145]

[0146] in, is the normalized data, D max is the maximum value of the training sample set, D min is the minimum value of the train...