MODELLING VOLATILITY AND FORECASTING BIST100 RETURN BY USING ANFIS

Stock market volatility is considered as an important issue in financial literature and is defined as sudden instability that occurs in the price of any security. Volatility also represents the uncertainty that affects investors' decision-making processes in financial markets in the face of possible variations. In many countries, especially in emerging financial markets, both investors and policy makers are often confronted with increasing risk and uncertainty problems. Accordingly, considering volatility is crucial for investors to predict the return of financial assets, especially in long-term investment decisions. Volatility, which expresses the variability of any financial asset, has a very important place in the estimation of the return. In this study, the volatility of the Turkish stock market was estimated through the GARCH models using the Stock Exchange Istanbul100 (BIST100) index. Whether the stock market index has an asymmetric effect is investigated using the EGARCH model. It is very difficult to predict the uncertainty and chaotic behavior of the BIST100 index by traditional methods. For this reason, the fuzzy logic and neural network hybrid model, which is widely used in the model of uncertainty, has been applied to estimate the stock return in the study. The dataset used in the study includes daily stock closing prices for the period 2009-2017. As a result of the extensive literature survey, no studies have been found on the use of fuzzy logic based approaches in estimating and continuing volatility. For this reason, it is thought that working has an original value