"[
ABSTRAKRegresi linier merupakan suatu metode yang dapat digunakan untuk memodelkan hubungan antara suatu variabel terikat terhadap satu atau lebih variabel penjelas. Terdapat beberapa asumsi yang harus dipenuhi pada model regresi linier, yaitu komponen error berdistribusi normal dengan mean nol, variansi error konstan
(homoskedastis), dan error antar observasi saling bebas. Pada saat menganalisis data spasial dengan menggunakan model regresi linier, asumsi homoskedastis terkadang tidak dapat terpenuhi karena kondisi data pada satu lokasi berbeda dengan kondisi data pada lokasi lainnya. Model Geographically Weighted Regression (GWR) dapat digunakan untuk mengatasi masalah heterogenitas spasial. Parameter model GWR dapat ditaksir dengan menggunakan dasar metode Weighted Least Squares (WLS) dimana bobotnya merupakan fungsi pembobot kernel. Fungsi pembobot kernel yang digunakan pada penelitian ini adalah fungsi pembobot kernel Gaussian. Pada bagian akhir penulisan, diberikan contoh aplikasi model GWR dengan menggunakan data klaim rawat inap peserta asuransi
kesehatan PT. XYZ untuk melihat hubungan antara total biaya rawat inap terhadap lamanya pasien dirawat inap dan kelas kamar rumah sakit yang ditempati pasien selama menjalani rawat inap untuk diagnosa tipes, DBD, dan diare. Dari hasil penelitian, hanya diagnosa tipes dan DBD yang dapat dianalisis dengan GWR. Berdasarkan peta penyebaran hasil taksiran parameter model GWR dan taksiran rata-rata total biaya rawat inap pasien dengan diagnosa tipes dan DBD, terlihat adanya variasi biaya di rumah sakit yang satu dengan rumah sakit lainnya.
ABSTRACTLinear regression is a method that can be used to model the relationship between a dependent variable to one or more explanatory variables. There are some assumptions that must be fulfilled in the linear regression model, such as the error term is normally distributed with mean zero, a constant variance (homoscedastic),
and independent among observations. When analyzing spatial data using a linear regression model, sometimes the homoscedastic assumption cannot be fulfilled because data condition on one location differ compared to others. Geographically Weighted Regression (GWR) model is used to overcome the spatial heterogeneity
problem. Parameters of GWR model can be estimated using Weighted Least Squares (WLS) method as the basic of estimating parameters using kernel weighting function. The kernel weighting function used here is Gaussian kernel weighting function. At the last chapter, there is an example of the GWR model application by using inpatient claims data of PT. XYZ members to see the relationship between the total inpatient cost to length of stay and hospital’s room rate for typhoid, DBD, and diarrhea. From the result, only typhoid and DBD that
can be analyzed with GWR model. Based on the map of parameter estimates on GWR model and average of total inpatient cost for typhoid and DBD, it shows that there is a variation of inpatient cost between one hospital and the others., Linear regression is a method that can be used to model the relationship between a
dependent variable to one or more explanatory variables. There are some
assumptions that must be fulfilled in the linear regression model, such as the error
term is normally distributed with mean zero, a constant variance (homoscedastic),
and independent among observations. When analyzing spatial data using a linear
regression model, sometimes the homoscedastic assumption cannot be fulfilled
because data condition on one location differ compared to others. Geographically
Weighted Regression (GWR) model is used to overcome the spatial heterogeneity
problem. Parameters of GWR model can be estimated using Weighted Least
Squares (WLS) method as the basic of estimating parameters using kernel
weighting function. The kernel weighting function used here is Gaussian kernel
weighting function. At the last chapter, there is an example of the GWR model
application by using inpatient claims data of PT. XYZ members to see the
relationship between the total inpatient cost to length of stay and hospital’s room
rate for typhoid, DBD, and diarrhea. From the result, only typhoid and DBD that
can be analyzed with GWR model. Based on the map of parameter estimates on
GWR model and average of total inpatient cost for typhoid and DBD, it shows
that there is a variation of inpatient cost between one hospital and the others.]"
Depok: Fakultas Matematika dan Ilmu Pengetahuan Alam Universitas Indonesia, 2015