Bankruptcy Prediction of Family Firms Using Combined Classifiers

Abstract

Literature has dealt with prediction of corporate bankruptcy since some decades ago, building classification methods to predict financial firm failure in different industries, countries and regions. However, and although the previous research is profuse, no paper in this literature has developed specific models for family firms. Family firms represent sixty percent of total firms in the EU and eighty percent in the USA, and it seems essential to evaluate their particular risk of bankruptcy. In this article we used combinations of classifiers (Naïve Bayes Classifier, Algorithm C4.5, Multilayers Perceptron and Support Vector Machine) through the AdaBoost algorithm, to develop an effective model to predict the bankruptcy of family firms.