ABSTRACT:

Due to a variety of influencing factors, the nature of stock market movement has long been unclear to investors. This work uses machine learning and deep learning techniques to significantly reduce the risk related to trend prediction. Four stock market groups from the Tehran Stock Exchange are selected for experimental evaluations: diversified financials, petroleum, non-metallic minerals, and basic metals. Nine machine learning models (Decision Tree, Random Forest, Adaptive Boosting, eXtreme Gradient Boosting, Support Vector Classifier, Naive Bayes, K-Nearest Neighbors, Logistic Regression, and Artificial Neural Network) and two potent deep learning techniques (Recurrent Neural Network and Long Short-Term Memory) are compared in this study. Our input values are 10 technical indicators from ten years of historical data, and two methods are intended for using them. First, stock trading values are used to calculate the indicators, and then, before use, the indicators are converted to binary data. Based on the input methods, each prediction model is assessed using three metrics. The evaluation findings show that RNN and LSTM perform significantly better than other prediction models for continuous data. Additionally, the results showed that those deep learning techniques are the best for evaluating binary data; however, the difference between them is decreasing as a result of the second method’s performance clearly rising.

Keywords: LSTM, ANN, STOCK MARKET