Malaysia Unemployment Rate Forecasting: Neural Networks Versus Box Jenkins Method

Abstract

The percentage of the labour force that is actively looking for work but is not yet employed is known as the unemployment rate. It is a crucial economic indicator that is used to evaluate the state of the labour market and the economy as a whole. There are three objectives in this study that has been achieved which are to analyse and predict the future value of unemployment rate data in Malaysia using ARIMA and ANN model. This study was focus more on last objective which is to differentiate univariate unemployment rate data using ARIMA model and multivariate unemployment rate using ANN model. The total number of observations from 2010 to July 2023 which make the data count is 163. The training data is from 2010 until 2022 while testing data is from January 2023 until July. The monthly statistics from January through December are the main emphasis of this report. Box-Jenkins of ARIMA model and Artificial Neural Network method are the two methods that have applied in this study. These approaches were implemented using Microsoft Excel, Minitab and SPSS software. The assessment of this unemployment rate was analysed in the discussion section and the results are presented in the result section which showed the forecast accuracy performance and the differentiation between the univariate unemployment rate data using ARIMA model and multivariate unemployment rate using ANN model.