A Comparison Between ARIMA and ANN in Predicting Population Growth in Malaysia

Soh, Zi Hui (2024) A Comparison Between ARIMA and ANN in Predicting Population Growth in Malaysia. Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text RMM_Soh Zi Hui.pdf Restricted to Registered users only Download (1MB)

Abstract

This study delves into the dynamic areas of forecasting the population growth rate in Malaysia from 1951 to 2023 by employing time series analysis and artificial neural networks(ANNs), aiming to examine and compare these methods. Time series analysis and ANNs forecast have been discussed deeply in this research. First and foremost, this project outlines the study’s background, problem statement, aims and objective as well as specifies its scope and significance of the study in Chapter 1. Then, the literature review in the second chapter encapsulates an exploration of relevant works in the field, providing a foundation for the methodologies selected. Chapter 3 investigates the detailed methodology, offering an introduction to both time series and ANNs, with particular emphasis on models selected to use in this study. Two models are investigated in this research which are the Auto-Regressive Integrated Moving Average Model(ARIMA) and Neural Network Autoregressive(NNAR). All results are presented and discussed in Chapter 4. Both ARIMA and NNAR models are rigorously fitted, compared and evaluated to determine the most appropriate forecasting method. The discussion revolves around the strengths and weaknesses of each model, elucidating their applicability in the context of Malaysian population growth. Chapter 5 culminates the study, providing a conclusive summary along with limitations and recommendations for future research. It highlights the achievement of the objectives outlined in the study's introduction. Keywords: population growth rate, time series analysis, ARIMA model, artificial neural network, NNAR model.