COMPARISON OF PARAMETRIC AND NONPARAMETRIC METHODS IN SEMIVARIOGRAM MODELING OF BUILT LAND COVER IN EAST JAVA

Semivariogram is a statistic used to measure the spatial correlation of pairs of locations with a certain distance and slope. Land cover data in the form of physical and economic factors can be modeled using spatial or geostatistical analysis. Physical factors include elevation, slope, and curvat...

Full description

Saved in:
Bibliographic Details
Main Author: Yeremy Budiman, Yonathan
Format: Theses
Language:Indonesia
Online Access:https://digilib.itb.ac.id/gdl/view/83087
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Institut Teknologi Bandung
Language: Indonesia
Description
Summary:Semivariogram is a statistic used to measure the spatial correlation of pairs of locations with a certain distance and slope. Land cover data in the form of physical and economic factors can be modeled using spatial or geostatistical analysis. Physical factors include elevation, slope, and curvature, while social factors include population density, distance to the nearest road, distance to the National Capital, distance to residential areas, and distance to the center of business activities. This research aims to build a semivariogram model based on land cover, especially built-up land, using parametric and nonparametric models. The least squares method will be used to estimate the parameters of the parametric semivariogram model, with the models chosen being the exponential and Gaussian models. For nonparametric semivariograms, the kernel functions used are Epanechnikov and Gaussian, with Naradaya-Watson kernel regression applied using NumXl software to obtain the semivariogram model. The modeling results show that the Gaussian model for parametric semivariograms has a smaller SSE compared to the exponential method. In contrast, for nonparametric semivariograms, the semivariogram using the Epanechnikov kernel has a smaller SSE than the nonparametric semivariogram using the Gaussian kernel.