Stock Price Prediction using Bi-directional Long-Short Term Memory based Deep Neural Network

In the area of finance, forecasting of stock price is becoming very popular increasingly. Predicting the stock price of a firm is very important to increase speculators interest in investing for a company and for growth of shareholders in that stock. In addition to a unique investment portfolio model based on the projections produced, this article suggests a novel method for time-series stock price prediction. To achieve this, a Bidirectional Long-Short Term Memory-based Deep Neural Network (Bi-LSTM-DNN) is equipped with novel regression technique. Once made, the predictions are utilized to build investment portfolio, more precisely, a projected portfolio. Numerous tests have been conducted using NIFTY-50 stock data that was acquired from Indian National Stock Exchange. The outcomes verify that suggested model performs better than a number of conventional prediction models and a number of conventional portfolio optimization algorithms. Reducing investing risks and increasing rewards are possible with an accurate stock price prediction. The Root Mean Square Error (RMSE) value for the proposed deep learning network for stock prediction using Bi-LSTM-DNN is 0.81.