Framing Qualitative Bankruptcy Prediction Rules Using Ant Colony Algorithm

Abstract

This paper is to frame qualitative Bankruptcy Prediction (BP) rules using the concept of Ant Colony Algorithm. There are various researches in the area of qualitative BP, among them Genetic Algorithm (GA) seems to more effective. But the redundancy and over lapping of the generated rules cannot be overcome by GA. In our work, we are proposing ACO for generating the rules for qualitative BP. The heuristic function and the pheromone trails are some of the features of ACO, which helps to give a positive feedback in generating rules and to avoid false rules. We are using Association Rule Miner (ARM) algorithm to cluster the generated rules and Partial Swarm Optimization (PSO) Technique to get the best rules among the generated rules. By using this we can generate better rules with more qualitative factors and redundancy and overlapping of the rules can also be avoided.