Inflation Convergence Modeling Using Binary Logistic Regression With SGD-Newton Raphson Optimization Methods in Indonesia

Fatma Novalia Kussumarani
Istiqomah, Nerissabila Uswatun
Siva Ifin Azzahra
Anggraini Puspita Sari
Sischa Wahyuning Tyas

EKSAKTA: Journal of Sciences and Data Analysis
VOLUME 7, ISSUE 1, April 2026

Global economic changes have necessitated the development of inflation models that can accurately describe Indonesia's economic dynamics. This study aims to compare two optimization methods, Newton Raphson and Stochastic Gradient Descent (SGD), in binary logistic regression modeling to analyze the effectiveness of monetary policy. This study contributes to evaluating the performance of both methods in terms of convergence speed and accuracy of inflation model parameter estimation. The results of the analysis show that the Newton Raphson method is more efficient in achieving convergence with an iteration value of 0.2933 compared to SGD, while both methods produce equivalent model quality based on the Akaike Information Criterion (AIC) values of 34.4008 and 34.4254. These findings emphasize the importance of selecting the right optimization method to support more efficient monetary policy analysis.