Top Highlights

• he reinforcement learning for stock price prediction hasnot yet received enough support as it should be. The main issue of supervisedlearning algorithms is that they are not adequate to deal with time-delayedreward [22, 18]
• In other words, supervised learning algorithms focus only onthe accuracy of the predictionat the momentwithout considering the delayedpenalty or reward. Furthermore, most supervised machine learning algorithmscan only provideaction recommendationon particular stocks1, using reinforce-ment learning can lead us directly to thedecision makings...
• There are two main applications of using machine learning in the stock markets:stock price prediction and stock trading.Stock price prediction can be divided into two applications: price regressionor stock trend prediction. In the first application, the researchers aim to predictexactly the numerical price, usually based on day-wise price [15] or c...
• 44]. Traditionally time-series forecastingtechniques such as ARIMA and its variant [43, 26] are adapted from the econo-metric literature. However, these methods cannot cope with non-stationary andnon-linearity nature of the stock market [2].
• Nevertheless, with the development of the machine learning community, tech-nical analysis gains attention of researchers in recent years.

Domain