A Portfolio Optimization Model Based on Information Entropy and Fuzzy Time Series

In this paper, a forecasting-mean-correlation-entropy portfolio optimization model (FMCE) is developed by using the fuzzy time series techniques to predict securities' future returns distribution and employing entropy as risk measurement. Traditional portfolio models such as MV model have stringent conditions to returns distribution, while entropy as a new risk measurement is free from these restrictions. Besides, traditional models assume securities' future returns distribution the same with their historical distribution which may be not suitable for the complex market. Therefore, a forecasting method is applied in FMCE. Based on the historical data of the Shanghai Stock Exchange and Shenzhen Stock Exchange in China, a comparison among the proposed model, hybrid models and traditional ones which are the mean-variance model, absolute deviation model and maximum absolute deviation model is made. The empirical results show that the proposed FMCE model is an effective tool in portfolio selection.