A CNN-based Stock Price Trend Prediction with Futures and Historical Price

In this paper, focusing on the task of feature extraction using financial time series as well as trend prediction for prices, a new stock sequence array convolutional neural network model is presented. This model shows promising results in improving the accuracy in stock trading forecasts. The implemented model collects data from historical sources and futures (leading indicators) of stocks. The model then uses arrays as the input map for a Convolutional Neural Network (CNN) framework. Through in-depth experimental results using the Taiwanese stock market, we are able to show that the designed approach achieves strong results. Furthermore, where compared with state-of-the-art similar models, our results show promising performance among all compared approaches.