Clustering Time Series Data by SOM for the Optimal Hedge Ratio Estimation

Abstract

The fat-tailed and leptokurtic properties observed in most financial asset return series would cause the inaccuracy of hedge ratio estimation because most traditional statistics approaches are based on the assumption of normal distribution. In this study, a novel approach is proposed using self-organizing map (SOM, also called Kohonen's Self-Organizing Feature Map) for time series data clustering and similar pattern recognition to improve the optimal hedge ratio (OHR) estimation. Five SOM-based models (considering the weight for averaging and the interval for data sampling) and two traditional models (ordinary least square method and naive hedge) were compared in Taiwan stock market hedging. The experiment demonstrates the feasibility of applying SOM, and the empirical results show that SOM approach provides a useful alternative to the OHR estimation.