Regresi Terboboti Geografis dengan Fungsi Pembobot Kernel Gaussian pada Kekuatan Sinyal Seluler

Authors

  • Logananta Puja Kusuma Departement of Statistics, IPB
  • . Indahwati Departement of Statistics, IPB
  • Kusman Sadik Departement of Statistics, IPB

DOI:

https://doi.org/10.29244/xplore.v8i1.134

Keywords:

cluster analysis, gaussian kernel, geographically weighted regression

Abstract

Cellular signal strength may be affected by its location, so researches concerning signal strength need information about location and analysis method that observe spatial aspect. Spatial Regression analysis evaluates location in modeling relation between explanatory variables and response variable. One of the spatial regression analyses is Geographically Weighted Regression (GWR). This method utilizes location to create weight matrix using certain weighting function. GWR analysis with Gaussian kernel weighting function creates better model than Ordinary Least Square model. The model created using GWR is local model which parameter estimation differs in each observation point. Clustering of observation point is performed to summarize the result of GWR. The number of optimum clusters in clustering based on coefficient is five clusters while the number of optimum clusters in clustering based on p value of t test is four clusters.

Published

2019-04-06

How to Cite

Kusuma, L. P., Indahwati, ., & Sadik, K. (2019). Regresi Terboboti Geografis dengan Fungsi Pembobot Kernel Gaussian pada Kekuatan Sinyal Seluler. Xplore: Journal of Statistics, 8(1). https://doi.org/10.29244/xplore.v8i1.134

Most read articles by the same author(s)