The Analysis on the Application of Machine Learning Algorithms in Risk Rating of P2P Online Loan Platforms

This paper introduces a machine learning algorithm to evaluate the risk level of the current P2P online lending platform in my country. This article is divided into two levels: firstly, estimate the risk of online loan platform; secondly, assess the credit risk of borrowers on the platform. Introduce the unsupervised learning algorithm in machine learning dichotomous k-means clustering, classify online lending platforms, analyze the performance of various platforms, determine the level of each category, and give the risk rating results of online lending platforms. Further use the supervised learning algorithm to study the credit risk of the borrower of the online lending platform. This paper uses the K-Means algorithm and related functions in AdaBoost to achieve data crawling of such dynamic web pages. After the end, it was found that there were still missing data on some platforms, and finally a large amount of information on the official website of the P2P online lending platform was manually checked to collect the required data. After collecting the data, the author processed the string by constructing a large number of regular expressions to correct obvious errors and provide data consistency. The experimental research results show that this article sorts the classification results by analyzing the variable values of the online loan platform to obtain different classification levels. In order to better rate the online loan platform, the rating results are more comprehensive and accurate.