From Documentation
Jump to: navigation, search

Deep learning has made breakthroughs in many areas, such as image recognition, speech recognition, language modeling, etc. In this seminar, we will introduce an RNN (Recurrent Neural Network) and demonstrate how to use it to predict trends in the stock market. First, we will focus on a special RNN – LSTM (Long Short-Term Memory) network, which is capable of keeping and making use of long memory of the past to forecast the future trend. Then, I will use S&P 500 index as time series data to show every step of the analysis: from data reformatting to creating a LSTM network, to training and prediction. NOTE: Our goal is to demonstrate how to build/train a LSTM model, not to describe how to develop a practical stock prediction system.