Design And Implementation Of P2P Net Loan Risk Analysis System Based On Android APP Content Dynamic Collection Method

Since the implementation of the reform and opening up,China's economy has developed rapidly,and private lending has embarked on the Internet stage,forming a new way of lending--P2P network lending.In the traditional financial field,the party that lends funds is a lending institution.At this time,the creditor is a lending institution.The platform is profitable through loan interest and bears the risk of loans.In the P2P online lending industry,this situation has changed,and the party that lends funds has changed.It becomes an individual.At this time,the lender becomes the creditor and bears the main transaction risk.The platform obtains the profit by extracting the transaction.The change of the profit mode causes the platform to shift its focus from trading risk control to facilitating more transactions,resulting in lower loan threshold.It has caused many social problems.In order to safeguard the tangible benefits of lenders,control the risk of online lending transactions,and ensure the security of online lending transactions,it is necessary to analyze the risks existing in online lending transactions.From the perspective of third parties,this paper analyzes the credit risk existing in online loan transactions and implements the online loan risk analysis system.The detailed work is as follows:1)In order to effectively analyze the risk of online loan transactions,it is necessary to obtain real and comprehensive transaction data.In recent years,with the increase in the number of mobile Internet users,the mobile terminal has become the main market for online loan platforms,and the same loan transaction platform will be on the mobile side.In order to disclose more information,this paper proposes a dynamic collection method for Android APP content,and starts the research data acquisition from the mobile terminal.2)In view of the credit risk problem existing in online loan transaction,it is necessary to analyze the online loan transaction data in detail and explore the difference between default loan and non-default loan.This paper first uses the correlation test to study the relationship between data features.It is found that the loan amount is highly correlated with the borrower's information quota and the monthly debt amount.This paper also uses the non-parametric test to study the difference between the default loan and the normal loan transaction,and finds the two types of borrowing amount,company size,and historical overdue.There are significant differences between the repayment period,the number of loan failures,and interest rates.In addition,by observing experimental data,this paper finds that the proportion of defaulted loans in loan transactions is extremely small,only 3%.3)Based on the above research,this paper aims at balancing the data set by using the undersampling method to reduce the proportion of the large sample in the training set.This paper uses a random forest algorithm that performs well on unbalanced data to construct a loan risk assessment model.The model evaluates the risk of the transaction by predicting the default probability of the borrower in the loan transaction.The higher the default probability,the more the transaction risk is.high.Compared with the existing literature,combined with the undersampling and integrated learning methods,the model effect is significantly improved.Finally,this paper designs and implements the P2P network loan risk analysis system,and realizes the application of the online loan risk assessment model with the concept of micro-service.