Forecasting Battles in Financial Time Series Data: Comparison of Traditional and Machine Learning Techniques

Abstract

Forecasting has an essential position for the improvement of any business. Financial markets are difficult to predict because many factors such as business risk, economic factors, political changes and Govt. policies etc. influence it. The objective of the study is to find the best forecasting method. The method could be from traditional or machine learning techniques. AI (Artificial Intelligence) is a significant technological advancement that has everyone buzzing about its incredible potential. The current research study evaluates the influence of artificial intelligence techniques i.e., Machine learning techniques and traditional techniques. In our study we have used data of 20 years (2000-2019) of cherat Cement Company traded on PSX. Different traditional and machine learning techniques have been used in the study. In this study, we have used stock historical prices for accurate stock price forecasting by developing machine learning algorithm using techniques namely Decision Tree, Neural Networks, (SVM) and traditional techniques like ARCH, GARCH and EGARCH. Secondary data collection technique has been used. Data is analyzed and tested through Python and Eviews software. Our results prove that AI techniques can accurately forecast stock prices with minimum error values and among all the techniques used Decision Tree perform better than rest of the techniques. The study will not only helpful for investors but also it will be helpful for brokers to forecast the stock prices.