Analysis of Financial Time Series Forecasting using Deep Learning Model

Abstract

Time series data analysis and its forecasting is a foremost trend of stock market prediction. Accurate prediction of stocks brings more profit to market traders and helps in financial decision making. There are various machine learning and deep learning models assist to predict the stock market accuracy. Recent work concludes that various models like Support Auto Regressive Integrated Moving Average (ARIMA), Vector Machine (SVM), Artificial Neural Network (ANN), XGBoost, and Recurrent Neural Network (RNN) were preferred to obtain improved accuracy. In this study, a stacked Long Short-Term Memory (LSTM) model is proposed to predict the stock market accuracy and proposed model is compared with Moving Average (MA) and XGBoost models. The experiments are performed on the historical dataset of Infosys Limited of Bombay Stock Exchange, India (BSE30). The model is also evaluated through performance measures Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) and found that proposed stacked LSTM model outperformed the benchmark model.