Genetic-based fuzzy models: Interest rate forecasting problem

Many phenomena in our lives are difficult to predict. Especially financial markets have eluded successful prediction attempts. Interest rates are quite volatile and nonlinear. We develop the system capable of processing Korean financial data and modeling time-series processes (such as interest rate) with fuzzy logic and genetic algorithms(GAs). In this paper, we bring together two technologies: fuzzy theory and genetic algorithms. The combination of these techniques could be applied to the interest rate forecasting problem in Korean financial market. The fuzzy rules can be concisely represented with one or more FAM (Fuzzy Associative Memory) matrices. We use GAs to adapt the FAM matrix entries so that the interest rate forecasting problem leads to an improved performance. This paper presents the Genetic-Based Fuzzy Model (GBFM).