Perbandingan Regresi B-Splines dan P-Splines pada Hubungan Indeks Pembangunan Manusia dan Persentase Penduduk Miskin Kabupaten/Kota di Indonesia

Poverty is basically a manifestation of the opportunity imbalance that every human being has, which occurs due to the imbalance of capabilities possessed. Measuring poverty by using dimensions of capability such as education, health and quality of life standards, can be a reference for identifying the characteristics of actual poverty. One of the measuring that takes into account the capability dimensions is the Human Development Index (HDI). Knowing the close relationship between HDI and the percentage of poor people, in this study a comparison of B-Splines and P-Splines nonparametric quantile regression methods was used to modeling the relationship between HDI and the percentage of poor population according to districts or city in Indonesia in 2017. The results of the analysis showed that the distribution data with curve fittings using non-parametric regression B-Splines and P-Splines produce smooth curves reaching all existing data distributions. The comparison of MSE B-Splines and P-splines models showed that the B-splines regression model for the HDI variable and the percentage of poor people gave the smallest MSE value of 0.1928378, so that it was the best model to analyze the relationship between HDI data and the percentage of poor people in district or city ​​in Indonesia in 2017.