Time Series Analysis for Stock Price in Main Market Bursa Malaysia using ARIMA and Artificial Neural Networks

Abstract

Stocks are a vital part of financial markets because they allow people and organizations to invest in company stock and also make significant contributions to stability and growth in the economy. For investors, financial analysts, and policymakers, accurate stock price forecasting is essential because it facilitates well-informed decision-making and efficient risk management in stock investment. This study investigates the performance of different time series models in forecasting the stock prices of a company listed on the Main Market of Bursa Malaysia such as Artificial Neural Network (ANN) and Autoregressive Integrated Moving Average (ARIMA) models. The objective of the study is to identify the most appropriate approach for stock price prediction among both ARIMA and ANN models. The historical monthly stock price data were obtained from Yahoo Finance from 2015 to 2024. The dataset was separated into training and testing sets for the purpose of developing and validating the model. To evaluate the accuracy of the model, evaluation metrics such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE) and Mean Absolute Percentage Error (MAPE) were used. The findings suggest that the ANN model demonstrated consistent performance with low error values and achieved higher accuracy compared to ARIMA. The MAPE of ANN in both testing and training sets is less than 10% which demonstrates high accuracy of forecasting. In conclusion, the results offered insightful information for stock prediction besides improving investment strategies and enhancing financial market stability.