Abstract

Gross Domestic Product (GDP) is a principal indicator used to measure the economiccondition of a country. Indonesia’s GDP growth from 2017 to 2019 was approximately6 percent; however, it experienced a decline in 2020 and 2021, with rates of only -0.02 percentand 2.41 percent, respectively. In the process of economic development planning, a forecastingsystem is required to determine GDP in the future. The forecasting method employed in thisresearch is fuzzy time series optimized using Particle Swarm Optimization (PSO), to enhancethe accuracy and convergence of forecasted values. The dataset used comprises secondary data,specifically 54 sets of Indonesian GDP data spanning from the first quarter of 2010 to the secondquarter of 2023. The analysis results indicate that the proposed method is better than the conventionalfuzzy time series approach. The former method provides a predictive value for one periodin the future with a Mean Absolute Percentage Error (MAPE) value of 4.40%. In contrast, thelatter yields higher predictive values with a MAPE value of 7.93%.

Keywords: Forecasting model; Gross domestic product; Fuzzy time series; Particle swarm optimization.