Spatial Autoregressive Quantile Regression untuk Pemodelan Angka Harapan Hidup di Provinsi Sumatera Utara

Life Expectancy (LE) is a key indicator for assessing the level of public health. North Sumatra Province still faces challenges in improving its LE, which remains below the national average. This study aims to compare three spatial models, namely: SAR, RSAR, and SARQR to model HLI in North Sumatra. Secondary data from 2023 for 33 districts/cities were obtained from the North Sumatra Provincial Statistics Agency, with HLI as the dependent variable and seven independent variables. The analysis began with multiple linear regression and classical assumption tests, followed by the formation of a spatial queen contiguity matrix, spatial effect tests (Moran's Index and Lagrange Multiplier), and modeling using the three spatial regression approaches. The results indicate that SARQR at the 0.1 quantile is the best model with the lowest AIC value (23.7764) and superiority in addressing spatial outliers. Therefore, SARQR at the 0.1 quantile is recommended as the optimal model for modeling AHH in North Sumatra.