Prediction of Stock Market Using an Ensemble Learning-based Intelligent Model

AI-based models have shown that stock market is predictable despite its uncertainty and fluctuating nature. Research in this field has further dealt with predicting the next step price amount and less attention has been paid to the prediction of the next movement of price. However, in practice, the necessary requisite for decision-making and use of the results of prediction lies in considering the predictable trend of stock movement along with predicting stock price. Considering the widespread search in the literature on the matter, this paper takes into account, for the first time, two criteria of direction and price simultaneously for the prediction of the stock price. The proposed model has two stages and is developed based on ensemble learning and meta-heuristic optimization algorithms. The first stage predicts the direction of the next price movement. At the second stage, such prediction and other input variables create a new training dataset and the stock price is predicted. At each stage, in order to optimize the results, genetic algorithm (GA) optimization and particle swarm optimization (PSO) are applied. Evaluation of the results, on the real data of stock price, indicates that the proposed model has higher accuracy than other models used in the literature.