A statistical inference framework for FSNBLR: Modeling underdeveloped regional status in Eastern Indonesia.

Area of Science:

• Socioeconomic modeling
• Regional development studies
• Statistical inference

Background:

• Persistent regional disparities in Eastern Indonesia highlight the need for advanced modeling to understand underdevelopment.
• Existing models may not fully capture complex nonlinear relationships influencing socioeconomic outcomes.

Purpose of the Study:

• To enhance the Fourier Series Nonparametric Binary Logistic Regression (FSNBLR) model with a statistical inference framework.
• To apply the enhanced FSNBLR model to identify determinants of underdevelopment in Eastern Indonesia.
• To compare the performance of FSNBLR against conventional Binary Logistic Regression (BLR).

Main Methods:

• Development of a statistical inference framework for FSNBLR, including simultaneous and partial hypothesis testing.
• Utilizing the Likelihood Ratio Test (LRT) for hypothesis testing.
• Application to data from 232 regencies in Eastern Indonesia (2021).

Main Results:

• Infrastructure quality and local fiscal capacity were identified as significant predictors of underdevelopment.
• The FSNBLR model demonstrated superior classification accuracy and lower Akaike Information Criterion (AIC) values compared to BLR.
• The model effectively captured nonlinear relationships among predictors.

Conclusions:

• The proposed inferential framework strengthens the FSNBLR model's foundation for complex binary response analyses.
• FSNBLR offers improved performance over BLR in socioeconomic studies, particularly in capturing nonlinearities.
• The findings provide valuable insights for addressing regional disparities and promoting development in Eastern Indonesia.