Forecasting unemployment rate in Malaysia: comparison between ARIMA and Fuzzy Time Series / Ahmad Faidhi Amir Faisol and Nur Azriani Mohamad Nor

• Main Authors: Amir Faisol, Ahmad Faidhi, Mohamad Nor, Nur Azriani
• Format: Book Section
• Language: English
• Published: College of Computing, Informatics and Media, UiTM Perlis 2023
• Subjects: Time-series analysis
• Online Access: https://ir.uitm.edu.my/id/eprint/100675/1/100675.pdf; https://ir.uitm.edu.my/id/eprint/100675/
• Institution: Universiti Teknologi Mara

According to Malaysia Labour Force Survey, the definition of unemployed is the person who was available for work but did not work during a reference period. Unemployment occurs when a person is available for work and actively looking for work but cannot find one. As unemployment is a gauge of economic health, a higher unemployment rate will negatively affect the labour market. In 2020, a new virus known as Coronavirus spread all through the world. According to the World Health Organization (2020), Covid-19 began as a localized health crisis but quickly became a global health crisis with severe economic consequences. Regarding Malaysia Informative Data Centre (MysIDC), the unemployment rate in Malaysia has rosily increased by 1.3% from 3.3% in 2019 to 4.6% in 2021. It will affect the country if it keeps increasing for the following year. Because of that, this study wanted to find the best model to forecast the unemployment rate. This study focused on the unemployment rate in Malaysia from 1982 to 2021. Two models; ARIMA and Fuzzy Time Series, will be used to determine which is better for forecasting by finding the minor error value. The result shows that the ARIMA (1,1,0) model better forecasts the unemployment rate than Fuzzy Time Series since it shows the smallest value for MAPE and MSE.