Prediction of economical investment tool returns with hidden Markov: In Turkey

Nowadays there is a great demand to financial investment tools due to the desire of profiting from future predictions. The revival in the markets drafted everyone interested in financial investment to information hunting. This causes an increase in future prediction studies. In this study BIST 100 index change rate is predicted by the Hidden MarkovModel using the exogenous factors affecting stocks. With the help of Hidden Markov Model's first solution algorithms, forward-backward algorithms, BIST 100 index change rate direction possibilities were predicted based on past values. Secondly prediction of the exogenous factor change affecting the BIST 100 index change rate was done and even though exact results were not reached, close predictions were done. Baum-Welch algorithm, which is a final solution algorithm, parameters were predicted again and found out to be very effective. Keywords: Hidden Markov model, Hidden Markov model algorithms, Borsa Istanbul, financial ınvestment tools