Application of Artificial Immune System in Constructing a Financial Early Warning System: An Example of Taiwanese Banking Industry

In recent decades, soft computing techniques have broadly applied to solve complex problems. Among the soft computing techniques, artificial immune system (AIS) have appeared as a new approach dealing with classification problems. In this paper, an AIS algorithm is developed and applied to a two-group classification problem. An example of Taiwanese banking industry is discussed and the financial ratios of each bank from 1998 to 2002 were collected. This system has to distinguish the operational performance (good or bad) of each bank to offer a reference material for the managers or investors. The performance of AIS is compared with other five early warning systems, namely, genetic neural networks (GNN), case-based reasoning (CBR), backpropagation neural network (BPN), logistic regression analysis (LR), and quadratic discriminant analysis (QDA). The result indicates that the proposed AIS is over 10% better than the three soft computing early warning systems (GNN, CBR and BPN). The AIS outperforms the statistical early warning systems (LR and QDA) at least 24%.