Price preduction in IMKB using neural networks

Abstract

The purpose of this thesis is to perform price prediction in Istanbul Stock Exchange (IMKB) using neural networks approach. The neural networks have been in use in the literature of plenty of time, however, this thesis is one of the first applications of neural network forecasting in the Turkish Financial Framework. The study focuses on four stocks, each of which exhibited different trends for the period january I 99 1 - June 1993. Comparative analysis were carried out for each prediction and detailed statistical inquiry was performed. Even though the full potential of neural networks could not be utilized (basically because of data limitations), the results prove that neural networks perform significantly successful predictions.