PERBANDINGAN MATRIKS PEMBOBOT ROOK DAN QUEEN CONTIGUITY DALAM ANALISIS SPATIAL AUTOREGRESSIVE MODEL (SAR) DAN SPATIAL ERROR MODEL (SEM)

The spatial weighting matrix is very important to overview of the relationship between one location to another in the spatial regression. In this study, the authors compare the weighting matrix of queen contiguity and rook contiguity in the SAR and SEM models in stunting cases in Bone Bolango Regency, Gorontalo Province. The variables used are the number of IDL, the percentage of LBW, the amount of proper sanitation, the percentage of exclusively breastfed babies, and the number of poor people. The purpose of this study was to determine the factors that influence stunting in Bone Bolango Regency, compare the results of the analysis of the rook contiguity and queen contiguity matrices in the SAR and SEM models and determine the best model and weighting matrix in stunting modeling in Bone Bolango Regency. The results showed that the significant factor in the SAR model was the number of poor people, while the significant factors in the SEM model were the number of IDL, the number of proper sanitation, and the percentage of exclusively breastfed babies. In the SEM model, the p-value of queen contiguity is smaller than that of rook contiguity.The best model in this study is the SEM model.