Prediction of IMKB sector indices by using artificial neural networks

Abstract

Stock market predictions play an important role for making right investment decisions. Investors can gain very high returns in short time in stock exchanges if correct stocks are chosen or they can lose their earning. Researchers are interested in stock markets for decades. Today, stock market topics are still examined by many scholars, since the factors determining the market conditions are changing continuously and none of the studies provide a complete and accurate solution for stock exchange direction. Academicians build many studies for modeling stock market behaviors and making different type of predictions such as selecting stocks with high rate of returns for portfolios, determining buy-sell point for indices or stocks, simulating economical crises time for providing alarm signals if crisis situations are likely, providing up or down signals for indices and etcetera. This study’s aim is to analyze the behaviors of sector indices of Istanbul Stock Exchange like XTRZM (Tourism companies), XKMY (Chemical companies) and make prediction for those sub-indices instead of making predictions for overall stanbul Menkul Kıymetler Borsası (IMKB) index. Artificial neural network approach which uses past data of sub-indices will be used to predict sector indices of IMKB. Stocks which constitute the sector index will be found and the stocks’ past data will be analyzed by artificial neural network approach.