The Validity of Using Technical Indicators When forecasting Stock Prices Using Deep Learning Models: Empirical Evidence Using Saudi Stocks

Abstract

Many researchers use deep learning and technical indicators to forecast future stock prices. There are several hundred technical indicators and each one of them has a number of parameters. Finding the optimal combination of indicators with their optimal parameter values is very challenging. The aim of this work is to study if there is any benefit of feeding deep learning models with technical indicators instead of only feeding them with price and volume. After all, technical indicators are just functions of price and volume. Empirical studies done in this work using Saudi stocks show that deep learning models can benefit from technical indicators only if the right combination of technical indicators together with their right parameter values are used. The experimental results show that the right combination of technical indicators can improve the forecasting accuracy of deep learning modules. They also showed that using the wrong combination of indicators is worse than using no indicator even if they were assigned the best parameter values.