Credit Card Risk Analysis Based On Petri Net And Process Mining

With the rapid development of China's economy,people's consumption awareness has changed,and the concept of "early consumption" has been rooted deeply in the people's mind.The personal credit business represented by credit card came into being,which greatly satisfied people's spiritual and material pursuit.The advantages of high quota and easy to use make people ignore its high risk nature.Scientific and reasonable credit card risk assessment can promote the healthy and vigorous development of credit card field.In view of the current situation that there is no causality assessment method for credit card risk in China,it is of great practical significance to study credit card risk analysis based on Petri net and process mining.In the past,the risk assessment of credit card often uses neural network method.The large number of neurons in the neural network makes the whole neural network can store a large number of effective information.Neural network is currently recognized as a classification model with high prediction accuracy.However,most of the previous studies directly used the indicator system for the over all qualitative analysis,lacking the description of the causal correlation of the development of credit risk events,which obviously has certain limitations.In order to better solve the above problems,this paper proposes an analysis method algorithm based on the combination of Petri net graphic modeling and fuzzy control theory.According to the specific rules,the fuzzy Petri net model of credit card risk is established,and BP neural network algorithm is introduced into the learning of model weight parameters.Finally,the effectiveness of the algorithm adopted in this paper is obtained through the analysis of specific experimental data.The main work of this paper is as follows:(1)For the problem of personal credit evaluation indicators,the existing research usually makes the indicators independent of each other,it's unable to reflect the interaction.A credit card risk assessment index system based on Fuzzy Petri net is proposed.In this paper,the index system of credit card risk assessment is decomposed by the method of decomposition process,and the indexes with common properties are classified into one category by using behavior outline,which solves the problem that credit card assessment indexes can not be qualitative,improves the causality between indexes,and reduces the error between indexes(2)Aiming at the problem of credit evaluation index weight in fuzzy Petri net,the existing literature mostly uses expert questionnaire method to get,the objective weight has used expert questionnaire,and the subjective weight continues to use this method to make the combination weight too subjective.For this problem,this chapter combines neural network and fuzzy Petri net,calculates the importance of each index by BP neural network,obtains subjective weight by using order relation method,and combines BP neural network and fuzzy Petri net to ensure the rationality and scientificity of fuzzy comprehensive evaluation model,so as to improve the rationality of credit evaluation value and convenience of subsequent calculation Calculate personal credit value.(3)In view of the problem that the indicators in credit evaluation can not be quantified after qualitative analysis,most of the existing methods,the data in personal credit database is used directly,nevertheless can not use the data information completely and reasonably.This paper constructs a credit card risk fuzzy comprehensive evaluation model based on entropy weight.In this paper,the entropy method is combined with the fuzzy mathematics theory,the risk classification level of the credit card security risk assessment index is taken as the comment,the membership degree is taken as the probability value,and then the entropy value of each index is calculated and weighted in turn,which greatly improves the objectivity,scientificity and rationality of the weighting.(4)In view of the shortcomings of large and incomplete credit evaluation information,a fuzzy Petri net credit card risk evaluation model based on entropy weight is proposed.Fuzzy Petri net has good parallel processing,rapidity and fault tolerance.It can process the data in credit evaluation model quickly and accurately Based on this,the feasibility of the fuzzy Petri net credit card risk assessment model based on entropy weight in improving the field of credit card risk analysis is verified by an example of credit card risk assessment model figure[23]table[10]reference[83]...