GENETIC-BASED FUZZY MODELS - INTEREST-RATE FORECASTING PROBLEM

Many phenomena in our lives are difficult to predict. Especially finan cial markets have eluded successful prediction attempts. Interest rate s are quite volatile and nonlinear. We develop the system capable of p rocessing Korean financial data and modeling time-series processes (su ch as interest rate) with fuzzy logic and genetic algorithms(GAs). Iri this paper, we bring together two technologies: fuzzy theory and gene tic algorithms. The combination of these techniques could be applied t o the interest rate forecasting problem in Korean financial market. Th e fuzzy rules can be concisely represented with one or more FAM (Fuzzy Associative Memory) matrices. We use GAs to adapt the FAM matrix entr ies so that the interest rate forecasting problem leads to an improved performance. This paper presents the Genetic-Based Fuzzy Model (GBFM) . (C) 1997 Elsevier Science Ltd.