Forecasting Consumer Price Index in Malaysia Through Time Series Analsyis

Yap, Yee Wen (2024) Forecasting Consumer Price Index in Malaysia Through Time Series Analsyis. Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text RMM_Yap Yee Wen.pdf Restricted to Registered users only Download (2MB)

Abstract

This research focuses on forecasting the Consumer Price Index (CPI) in Malaysia through Time Series Analysis. It utilizes annual CPI data from 1960 to 2022 and monthly data from January 2010 to December 2023. The study includes an analysis of the background, problem statement, aims and objectives, project scope, and significance. Additionally, it reviews a number of research papers and journals to gain a clear understanding of analysis techniques and forecast methods used in predicting the CPI. In the methodology section, five types of time series models are introduced for investigating CPI forecasting: Autoregressive Integrated Moving Average Model (ARIMA), Simple Exponential Smoothing Model, Holt’s Model, Holt-Winters Additive Model, and Prophet. Additionally, the research outlines the statistical tests used in the forecasting process. Several accuracy measurements will be conducted in this study to determine the best-fit model for forecasting both the annual and monthly CPI in Malaysia. These measurements are essential for selecting the most appropriate model that can accurately capture the underlying patterns and trends in the CPI data. The conclusion of this study highlights Holt's Model as the best-fit model for forecasting both the annual and monthly CPI in Malaysia. Some recommendations aim to enhance the forecasting capabilities of CPI forecasts has discussed. Keywords: Consumer Price Index (CPI), Autoregressive Integrated Moving Average (ARIMA), Exponential Smoothing Model, Prophet