The use of artifical neural networks in prediction of stock return: An application in Istanbul Stock Exchange

Artificial Neural Networks which is a nonlinear statistical model is used to solve the sophisticated and heavy costly problems today. When studies in the literature are examined the ANN model panned out more successful result in order to forecast stock returns. Macro-economic indicators and financial values of firms which indicate stock market performance of firms are used to predicting the earnings. However the numbers of studies in which financial ratios, economic indicators and stock market performance are used together are not adequate. In this study which aims to predict stock returns of manufacturing companies in ISE the data was used belongs to firms are trading since 1986. However, it is not possible to reach the data for ISE before 1991, 1991-2010 periods was selected as the analysis period. In the study, changes in common stock returns for years 2008, 2009 and 2010 classified by ANN model and measured the performance of model by comparing with actual values in the same year. Thus, the expected returns of common stocks have been estimated for the year 2011.