Assessment of Financial Risk Prediction Models with Multi-criteria Decision Making Methods

Abstract

A wide range of classification models have been explored for financial risk prediction, but conclusions on which technique behaves better may vary when different performance evaluation measures are employed. Accordingly, this paper proposes the use of multiple criteria decision making tools in order to give a ranking of algorithms. More specifically, the selection of the most appropriate credit risk prediction method is here modeled as a multi-criteria decision making problem that involves a number of performance measures (criteria) and classification techniques (alternatives). An empirical study is carried out to evaluate the performance of ten algorithms over six real-life credit risk data sets. The results reveal that the use of a unique performance measure may lead to unreliable conclusions, whereas this situation can be overcome by the application of multi-criteria decision making techniques.