Nuha Nurfauziyyah, Yuliana Susanti, Isnandar Slamet

Pemodelan Regresi Robust Metode Estimasi Generalized Maximum Likelihood pada Negara Low Human Development

Introduction

Pemodelan regresi robust metode estimasi generalized maximum likelihood pada negara low human development. Modelkan IPM negara low human development (2022) dengan regresi robust estimasi GM untuk atasi pencilan. Temukan faktor penentu IPM: harapan hidup, lama sekolah, & PNB per kapita.

Abstract

Indeks Pembangunan Manusia (IPM) dibentuk sebagai parameter pembangunan suatu negara. Perhatian terhadap negara-negara yang termasuk ke dalam low human development sangat penting untuk memastikan tidak ada satu pun negara yang tertinggal dalam pencapaian Sustainable Development Goals (SDGs) tahun 2030. Tujuan penelitian ini adalah untuk memodelkan IPM pada kelompok negara low human development tahun 2022 menggunakan regresi robust estimasi GM. Perhitungan model regresi robust dilakukan untuk mengetahui faktor yang mempengaruhi IPM dengan faktor terdiri atas umur harapan hidup, harapan lama sekolah, rata-rata lama sekolah, dan pendapatan nasional bruto per kapita. Model tersebut digunakan karena data mengandung pencilan sehingga sisaan melanggar asumsi normalitas pada model regresi. Oleh karena itu, penelitian ini menggunakan regresi robust yang tahan terhadap adanya pencilan. Hasil penelitian didapatkan bahwa regresi robust estimasi GM memiliki nilai adjusted R-squared, yaitu 98,467% dengan Akaike Information Criterion (AIC) sebesar -227,149.

Review

This study addresses a highly relevant issue by modeling the Human Development Index (IPM) in low human development countries, aligning with the critical objective of achieving the Sustainable Development Goals (SDGs) by 2030. The authors appropriately identify a methodological challenge: the presence of outliers and the resulting violation of normality assumptions in standard regression models. To overcome this, the research judiciously employs robust regression with Generalized Maximum Likelihood (GM) estimation, a sound approach for data exhibiting such characteristics. The independent variables selected—life expectancy, expected years of schooling, mean years of schooling, and GNI per capita—are well-established proxies for human development and provide a comprehensive framework for analysis. A key strength of the research lies in its robust methodological choice, which ensures the reliability of the model in the presence of potentially influential data points. The reported model performance metrics are highly encouraging, with an impressive adjusted R-squared value of 98.467%, indicating that the chosen factors explain a very high proportion of the variance in IPM among these countries. The inclusion of the Akaike Information Criterion (AIC) further suggests a consideration for model parsimony and fit. By successfully identifying factors influencing IPM, the study lays a crucial groundwork for understanding development dynamics in the most vulnerable nations. While the abstract effectively outlines the motivation and methodological approach, its utility for an expert audience could be significantly enhanced by providing more substantive details of the findings. Specifically, the abstract states the study's aim is "untuk mengetahui faktor yang mempengaruhi IPM," but it does not present *which* factors were found to be significant, nor their direction or magnitude of influence. This omission prevents a full understanding of the study's empirical contribution. Furthermore, a brief discussion on the specific type of GM estimator or weighting function used, alongside a comparison (even qualitative) of the robust model's outcomes versus a traditional OLS approach, would further highlight the benefits of the chosen method. Finally, linking the identified influencing factors directly to potential policy implications for achieving the SDGs would bolster the practical relevance of this otherwise well-conceived research.

Full Text

Comments

You need to be logged in to post a comment.