Investigation of the relationship between bist100 index and commodities, stock market prices and exchange indexes of developing countries using machine learning and deep learning method

The effect of foreign exchange markets, commodity markets and stock market indices of developing countries on Bist100 is very important. The economies of the countries are strongly dependent on the economies of both their own and other countries, and therefore the markets are affected. Forecasting methods are used to follow the market economically and to make correct decisions. In this study, using machine learning and deep learning methods, the relationship between Bist100 index and commodity, foreign exchange prices and stock market indices of developing countries was examined using estimation algorithms. The dataset consists of closing data between January 2017 and October 2021. In order to ensure objectivity in experimental studies, a k-fold cross-validation model was applied. In comparison of models; Mean Absolute Error (MAE), Relative Absolute Error (RAE), Root Mean Square Error (RMSE) were used. While IBk, Kstar, Random Committee and Random Forest models are used in machine learning, the LSTM model is used as a deep learning method. As a result of experimental studies, it has been seen that the LSTM model gives better results than machine learning models. When the test results for the LSTM model are examined, the MAE value is 10.27, the RMSE value is 14.15, and the RAE value is 6.06.