Prediction of TCS Stock Prices Using Deep Learning Models

A precise forecast is surely critical in stock prediction. Many people are increasingly investing in stocks in recent times; hence it is evident that prediction of stock prices is very crucial. Statistical, Machine Learning and Deep Learning models can be used to forecast Stock prices. In this paper, Statistical method SARIMA is used to predict stock prices. To improve the outcome of forecasting five Deep Learning models namely Single-Layer LSTM, Three-Layer LSTM, Bidirectional LSTM, CNN LSTM and Convolutional LSTM are used. In Single-Layer LSTM and Three-Layer LSTM models LSTM layers are used to give better output. Bidirectional LSTM is used to predict the prices not only based on the previous inputs but also upon the future inputs. CNN is combined with LSTM to capture the underlying abstract features in the input series. Convolutional LSTM is another variation of LSTM which helps to capture the features that contain convolution activity inside the LSTM cell. All the proposed models are validated using evaluation metrics like RMS, RMAE, MAE, R²-SCORE to determine the robustness of the every Deep Learning model used. For this paper, TCS stocks are considered. The proposed methods in this paper can be applied to any other company stocks for the forecasting their stock prices.