Modelling stock market exchange by autoregressive integrated moving average, multiple linear regression and neural network

Abstract

Stocks, sometimes known as equities, are fractional ownership shares in a firm, and the stock market is a venue where investors may purchase and sell these investible assets. Because it allows enterprises to quickly get funds from the public, a well-functioning stock market is critical to economic progress. The purpose of this study is to model Bursa Malaysia using autoregressive integrated moving average (ARIMA), multiple linear regression (MLR), and neural network (NN) model. To compare the modelling accuracy of these models for intraday trading, root mean square error (RMSE) and mean absolute percentage error (MAPE) as well as graphical plot will be used. From the results obtained from these three methods, the NN model provides the best trade signal.