PREDICTION OF INSOLVENCY IN NON-LIFE INSURANCE COMPANIES USING SUPPORT VECTOR MACHINES, GENETIC ALGORITHMS AND SIMULATED ANNEALING

In this paper we propose an approach to predict insolvency of non-life insurance companies based on the application of Support Vector Machines (SVMs), hybridized with two global search heuristics: a Genetic Algorithm (GA) and a Simulated Annealing (SA). A SVM is used to classify firms as failed or non-failed, whereas a GA and a SA are used to perform on-line feature selection in the ratios space of the SVM, in order to improve its perfor-mance. We use general financial ratios and also other specific ratios which have been proposed for evaluating insolvency of insurance sector. In the simulations section, we compare the performance of the GA and SA as part of the proposed algorithm. The results obtained with both techniques show that the proposed algorithm can be a useful tool for parties interested in evaluating insolvency of non-life insurance firms.