Data Analysis and Related Applications 3

2Multivariate Kernel Discrimination Applied to Bank Loan Classification

The purpose of this chapter is to apply a kernel discriminant analysis to classify bank loans and determine which loans are at risk of default. This study begins by introducing the concept of kernel density estimation, which is a widely used non-parametric technique to obtain an estimate for the probability density function. This procedure is based on two main parameters: the kernel function and the bandwidth, the latter being the crucial parameter. The multivariate kernel density estimator is later applied to discriminant analysis to obtain kernel discrimination. This is a method that classifies observations into a predetermined number of distinct and disjointed classes. Finally, we apply multivariate kernel discriminant analysis to a sample of bank loans to determine which loans can be classified as defaulted. This model can help predict the likelihood that future loans may default.

2.1. Introduction

As the name implies, credit risk analysis is used to assess the likelihood of a borrower’s repayment failure and the loss caused to the financer when default occurs. This concept greatly affects the long-term success of any bank or financial institution. In Malta, the non-performing loan ratio stood at 3% in December 2019 (Central Bank of Malta). This ratio reflects the country’s credit quality of loans. Banks need to determine the probability of non-performing loans of companies to decrease the prospect ...