<![CDATA[The Consumer Price Index (CPI) is a crucial macroeconomic indicator reflecting regional inflationary pressures. This study aims to compare the performance of Ordinary Least Squares (OLS) and Maximum Likelihood Estimation (MLE) methods in estimating Holt-Winters model parameters for forecasting the CPI of East Nusa Tenggara (NTT) Province. Using monthly data from January 2021 to December 2025, the research evaluates two primary approaches: conventional OLS based Holt-Winters models and the MLE based Error, Trend, and Seasonal (ETS) framework. Model performance was assessed using the Mean Absolute Percentage Error (MAPE). The results demonstrate that the MLE based approach significantly outperforms OLS; the multiplicative ETS model achieved the lowest MAPE of 0.57%, proving far more accurate than the OLS based Holt-Winters models, which yielded error rates exceeding 50%. A 12-month forecast for 2026 projects NTT CPI to increase moderately from 107.38 to approximately 109.54, indicating controlled inflationary pressure. This study affirms the superiority of the MLE approach in generating precise parameters for economic data exhibiting seasonal patterns.]]>
Copyrights © 2026
