A Weighted Voting Algorithm for Predicting Corporate Bankruptcy

초록

This study proposes a weighted voting approach to predict corporate bankruptcy more accurately, which combines statistical methods with artificial intelligence techniques. In particular, by integrating multiple discriminant analysis, logistic regression, artificial neural networks, rule induction, and Bayesian networks, we develop a voted convergence method and a weighted voting system using artificial neural networks. Financial data from the Korea Credit Guarantee Fund are used in the analysis. The experiment results show that the Weighted Voting System might be a promising approach to estimate the risk of corporate bankruptcy compared to the currently-in-use approaches.

키워드