Stacked Deep Learning Structure with Bidirectional Long-Short Term Memory for Stock Market Prediction

Ying Xu, Lymeng Chhim, Bingxin Zheng, Yusuke Nojima

Abstract: The rapid growth of deep learning research has introduced numerous methods to solve real-world applications. In the financial market, the stock price prediction is one of the most challenging topics. This paper presents design and implementation of a stacked system to predict the stock price of the next day. This approach is a method that considers the historical data of the real stock prices from Yahoo Finance. This model uses the wavelet transform technique to reduce the noise of market data, and stacked auto-encoder to filter unimportant features from preprocessed data. Finally, recurrent neural network and bidirectional long-short term memory are used to predict the future stock price. We evaluate our model by analyzing the performance of different models with time series evaluation criteria.