Stock Price Forecasting Using Machine Learning Techniques


Creative Commons License

KOÇ USTALI N., Tosun N., TOSUN Ö.

ESKISEHIR OSMANGAZI UNIVERSITESI IIBF DERGISI-ESKISEHIR OSMANGAZI UNIVERSITY JOURNAL OF ECONOMICS AND ADMINISTRATIVE SCIENCES, vol.16, no.1, pp.1-16, 2021 (ESCI) identifier

Abstract

This study aims to estimate the future prices of stocks of firms listed in Borsa Istanbul Joint Stock Company (BIST) 30 Index. For this purpose, firstly, quarterly financial statements of BIST 30 Index companies between 2010-2019 have been provided and then financial ratios of firms have been calculated through these tables. In addition, monthly closing prices of company stocks were reached, and quarterly averages were taken in line with the financial ratios of firms. After obtaining the data, the future price of each company's stock was estimated by using Artificial Neural Networks (ANN), Random Forest (RF) algorithm and XGBoost algorithm. Then, the estimation results obtained according to each method were compared. It was determined that although XGBoost and Random Forest algorithms gave similar results, XGBoost has slightly better forecast results. Also, both models performed better than ANN.