An Analysis of GRU-LSTM Hybrid Deep Learning Models for Stock Price Prediction

Abstract

Investment and national policy researchers are studying stock price forecasting, which has proven to be a challenging problem given the multi-noise, nonlinearity, high-frequency, and chaotic nature of stocks. Most forecasting models will not be successful in mining actual data from stocks if these characteristics are present. Stock pricing data has the characteristics of time series. It is evident from different studies that deep learning models perform better than machine learning models on time series data in particular. So, in this paper, we will focus on Long Short Term Memory (LSTM), Gated Recurrent Unit (GRU), and a hybrid model of them to predict the price of HDFCBANK stock. The first hidden layer is GRU and the other three hidden layers of LSTM. A hybrid model is validated using MSE, RMSE, and MAE and it outperforms all other models.