Assessing Unemployment Rate Forecasting Accuracy During COVID-19 Using Machine Learning

Unemployment is one of the components in macroeconomics that must be considered and maintained by the government to maintain the country’s welfare. Therefore, a model is needed to forecast the unemployment rate in the future in the hope that the government can prepare policies to reduce the impact of unemployment. This study used the time series analysis method with the Autoregressive Integrated Moving Average with Exogenous Variables (ARIMAX) technique. In this study, researchers used the unemployment rate dataset from 1991-2021, and they considered the dataset before and after the COVID-19 outbreak. Based on the research results, the ARIMAX model produces a reasonably accurate model of the dataset before the COVID-19 outbreak. However, the ARIMAX model generated from using the dataset after the COVID-19 outbreak is less precise.