PREDICTION OF THE RELATIONSHIP BETWEEN THE BIST 100 INDEX AND ADVANCED STOCK MARKET INDICES USING ARTIFICIAL NEURAL NETWORK: (2011-2015)

Authors : Kemal ADEM, Numan ZENGİN, Mahmut HEKİM, Süleyman Serdar KARACA

Pages : 86-95

View : 18 | Download : 7

Publication Date : 2016-09-01

Article Type : Research Paper

Abstract :Prediction techniques and models are significant for people and organizations who wish to make prediction at the stage of investment and decision-making. For investors who want to achieve high earnings from investments, the stock market indexes are extremely important. Price movements in the stock such as political and social ones are affected by many factors. In the studies conducted on İstanbul Stock Exchange Index insert ignore into journalissuearticles values(BIST-100);, the estimation generally of foreign exchange rates, interest rates, gold prices, GNP insert ignore into journalissuearticles values(Gross Nation Product);, CPI insert ignore into journalissuearticles values(Consumer Price Index); and the relationship with macroeconomic variables such as the rate regard to traditional statistical prediction models were used. In this study, international advanced BIST100 Index of the estimation with Artificial Neural Network insert ignore into journalissuearticles values(ANN); method will be used as input instead of traditional macroeconomic variables insert ignore into journalissuearticles values(independent variables); and also stock market index data sets will be used. From January 2011 to December 2015 period, daily closing price of some international advanced stock market indices and BIST 100 Index data were used as data set. Data analysis were carried out through Multilayer Neural Network insert ignore into journalissuearticles values(MLNN); method, which is an ANN model widely used in MATLAB and the successful rate was %96,92
Keywords : Investment, BIST 100, International stock market indices, Artificial neural networks, Prediction