A SIMPLE NEURO FUZZY MODEL FOR ISE 100 INDEX PREDICTION

16463

In this paper, Istanbul Stock Exchange (ISE) National 100 Index retrospective predictability was tested by using the Adaptive Neuro Fuzzy Inference System (ANFIS). In addition, contribution of the inputs to the model for index forecasting was evaluated on the basis of prediction performance. The most important factor to successful stock market prediction is achieving best results using minimum required input data and the least complex stock market model. In this context, we aim to prove consistent prediction of ISE 100 index without having to use a lot of input variables with ANFIS. For this purpose, about four and a half year period was chosen as the analysis period; two input variables (the exchange rate and repurchasing interest rate) and three input variables (the exchange rate, repurchasing interest rate and trading volume) were established in two different models. Consistent prediction results have been obtained with high predictive factor with both of the models. As a result, that is concluded that the ISE 100 Index has short-term predictability using only two input variables, without the need for a complex model with ANFIS

897

ISE 100 index prediction , ANFIS , neuro , fuzzy systems

Journal Of The Faculty Of Engineering an‎d Architecture Of Gazi University

904

Journal Of The Faculty Of Engineering an‎d Architecture Of Gazi University