MODELING OF ANISOTROPIC SEMIVARIOGRAM USING WEIGHTED LEAST SQUARES AS PARAMETER ESTIMATION METHOD (CASE STUDY: COVID-19 CASES IN BANDUNG CITY)

• Main Author: Vidyaningtyas Sabila, Fatsa
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/65350
• Institution: Institut Teknologi Bandung

Anisotropic semivariogram is a statistic used to measure spatial correlation of pair of locations that separated by certain distance and angle. Coronavirus Disease (COVID-19) is an infectious disease caused by the Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-Cov-2). The data of COVID-19 daily cases is an example of data that can be modelled spatially. This research goals are to construct anisotropic semivariogram model based on COVID-19 confirmed Active and Death cases of 30 districts in Bandung City and using this model to predict unobserved locations by Ordinary Kriging interpolation method. Weighted Least Squares is used for parameter estimation method of semivariogram model. The data is modelled by using theoritical semivariogram models such as Gauss, Exponential, and Spherical. The direction sectors used in anisotropic semivariogram modelling are East-West, Northeast-Southwest, North-South, and Southeast-Northwest. Jackknife Kriging method is used to determine the best model of each data. The best model for Active data is Geometry Spherical and the best model for Death data is Zonal Spherical. Sum Squares Error (SSE) of Jackknife Kriging for Active model is 2310.3710, while SSE of Jackknife Kriging for the Death model is 15487.3222, larger than Active model.