MULTINOMIAL LOGISTIC REGRESSION MODEL USING MAXIMUM LIKELIHOOD APPROACH AND BAYES METHOD ON INDONESIA'S ECONOMIC GROWTH PRE TO POST COVID-19 PANDEMIC

Abstract

Indonesia's economic growth is a major concern in the global context, especially before and after the Covid-19 pandemic. Key sectors such as tourism, manufacturing, trade, and transportation have been severely impacted by travel and economic activity restrictions imposed to control the spread of the virus. Therefore, modeling is needed to describe the existing conditions. In this study, two approaches were used, namely the Maximum Likelihood approach and the Bayes approach. The use of methods in general as research material for researchers to further study the two methods. So far, the algorithm used for the Bayes concept method is Markov Chain Monte Carlo with Hasting's Metropolis method. The parameter estimation results obtained from the two methods are considered quite identical. However, it is necessary to pay attention to the iteration procedure that will be carried out. The selection of factors used in the iteration process is very important in obtaining the estimated parameter values. Furthermore, the results obtained so far do not contain fundamental differences regarding Indonesia's economic growth. In general, Indonesia can be said to be stable in terms of economic growth.