Comparative analysis of stock price prediction between machine learning and time series models

Predicting the future direction of an increase or decrease in the stock prices of the shares of the companies listed on the stock exchange has been one of the problems discussed in financial terms. In this study, forecasting on the direction of price prediction was made using the closing values of 8 different companies in the Finance, Communication, Transportation, Consumption, Energy, Construction, Electricity-Water-Gas and Retail sectors in Borsa Istanbul between May 24, 2021 and May 24, 2023. Arima, Polynomial regression, Support Vector Machines and Artificial neural networks were used in the predictions. The forecasting results are evaluated by comparing them with a representative random forecasting method, the coin toss model.