Implementation of Long Short-Term Memory and Gated Recurrent Units on grouped time-series data to predict stock prices accurately

Abstract Stocks are an attractive investment option because they can generate large profits compared to other businesses. The movement of stock price patterns in the capital market is very dynamic. Therefore, accurate data modeling needed forecast prices with a low error rate. Forecasting models using Deep Learning believed be able predict movements accurately time-series input, especially Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) algorithms. Unfortunately, several previous studies investigations LSTM/GRU implementation have not yielded convincing performance results. This paper proposes eight new architectural for forecasting by identifying joint market. technique combine LSTM GRU four neural network block architectures. Then, proposed model evaluated three accuracy measures obtained from loss function Mean Absolute Percentage Error (MAPE), Root Squared (RMSPE), Rooted Dimensional (RMDPE). accuracies, MAPE, RMSPE, RMDPE, represent lower accuracy, true higher model.