The Analysis Of Profitability In Peer To Peer Lending Market Based On Maltiple Linear Regression

Past studies have developed P2P credit scoring,although no one has proposed a profit scoring.The profit Scoring is used to choose the borrow-er who is profitable the most,which is the age value of the customer.This thesis uses a multiple linear regression to analyze the impact of profitability in peer to peer lending market.The model is based on four variable sets,purpose of the loan,housing situation,credit history and borrower's indebtedness.Data collected from Lending Club(152,167)p2p loans from 2008 to 2014 are examined in this thesis.The P2P loan is a new type of loan where the lender and the borrower can meet and achieve the best deal on a common platform.Although borrowers look for lenders that give funds at lower interest rates,lenders look for borrowers with zero or lowest probability of default.We proposed to use Net Present Value of loan as the profitability measure,which captures all other forms of repayments to give a clear picture of the analysis.Unlike previous studies that focused on default probability in peer to peer lending,the target of this thesis is to predict the expected profitability of investing in p2p lending using multiple linear regressions and this research shows that some customers with a high probability of default may be profitable.Estimates of the size and statistical significance of relationships between variables have been provided.In specific,if the loan allocation is dictated by the credit scoring system designed to predict Probability of Default,the riskiest borrowers may not be able to get credit,though they may be useful and profitable.Our results show that main factors explaining the profitability in peer-to-peer are credit card,debt consolidation and revolving utilization loans.This new application of data mining techniques in profit scoring systems could extend remarkable improvements of profitability in p2p lending markets.Presentation of our analysis has been given in different tables and graphics.Machine learning algorithms are filling our everyday life.In profit scoring systems,linear regression has potential to classify and regress.Different companies are so soon using linear regression method to accelerate the loan process.According to Jeremy Liew's article in Bank of America,companies such as Zestcash and Global Analytics are becoming larger and faster by implementing machine learning techniques and big data in their systems.It enables borrowers to easily and quickly access loans at a lower interest rate.Linear regression is a statistical process that is usually used in different fields of science to define the connection among variables in the form of an equation to evaluate the power and way of the parameters' relationships.