Automatic Selection of Financial Ratios by Means of Differential Evolution and for Predicting Business Insolvency

Abstract

[Abstract]: Differential evolution was used for the automatic selection of financial ratios for the prediction of business insolvency. A sample of companies from the Galician economy is used to predict the insolvency of a company a year in advance. The genetic population encodes possible sets or combinations of financial ratios, and the quality of each encoded solution is determined by the classification accuracy provided by a KNN classifier that uses the selected encoded ratios. Finally, the selected and relevant ratios are used with a more robust classifier, a classical multilayer perceptron, which provides greater sensitivity in the prediction results.