Paper Title
Personal Loan Fraud Detection Based on Hybrid Supervised and Unsupervised Learning

Abstract
Abstract - In recent years, have been witnessing a dramatic increase on the personal loan for consumption, due to the quick growth of e-services, including e-commerce, e finance and mobile payments. Leads to the network's lack of effective inspection and supervision, which will inevitably lead to large-scale losses due to loan fraud. Considering how difficult it is to manually check and verify large numbers of credit card transactions, machine learning should be widely used to automatically detect fraudulent transactions. To filter unnecessary information and preserve useful information without knowing the meaning of data, this paper combines Kernel Principal Component Analysis (Kernel PCA) with XGBoost (algorithm) and proposes a supervised and unsupervised associative learning model, KPXGBoost. It uses network research to limit over-assembly and considers the exposure of XGBoost and PXGBoost and other traditional AI techniques. customer. Keywords - Supervised Learning, Unsupervised Learning, Extreme Gradient Bo Boost, Principal Component Analysis.