RONNY KOESNARIYANTO, 101041037 (2012) PEMODELAN INDIKATOR PENCEMARAN AIR SECARA KIMIA (BOD) DENGAN GEOGRAPHICALLY WEIGHTED REGRESSION (GWR) ( Studi Analisis Pemantauan Pencemaran Air Badan Air Sungai di Surabaya Pada Data BBTKLPP Surabaya Tahun 2011 ). Thesis thesis, UNIVERSITAS AIRLANGGA.

Preview	Text (ABSTRAK) gdlhub-gdl-s2-2013-koesnariya-23086-8.absta-t.pdf Download (203kB) | Preview
	Text (FULL TEXT) gdlhub-gdl-s2-2013-koesnariya-23086-23full.pdf Restricted to Registered users only Download (2MB) | Request a copy

Abstract

Spatial data is measurement data of location information. On spatial data, an observation in a location frequently depends on neighboring locations. Cressie (1993) stated that spatial data are one of types of dependent data because the data are collected from different spatial locations that indicate the dependence between data measurement and location. To overcome the problem of spatial data, the statistical method that will be used is Geographically Weighted Regression (GWR), namely a model that uses geographical factors as a predictor variable that can affect response variables. The indicators of body water pollution of water river of BOD values from rivers in Surabaya is closely related to location or position of the river point because each river has the characteristics of a river network such as width of a river, depth of a river, river flow rate, stream flow, and temperature of a river that are different from one another. The analysis to know the significant on BOD values and to get the best model is done with the analysis of GWR regression by using weighting kernel bisquare functions. The application of GWR model on the data of BOD values of rivers in Surabaya shows geographical factors or spatial variations that influence significantly. In other words, there is a significant difference between OLS and GWR models. Based on significant variables, rivers in Surabaya can be grouped into eight river groups that have the similar characteristics. The GWR model produces R2 value of 90,90% that tends to be larger, and SSE tends to be smaller with the value 30,78961 compares to OLS model with the R2 value = 69,91% and SSE 101,85050. In short, the modeling study of the water pollution indicators is that GWR model is better and appropriate to be used in the real world than OLS model.

Actions (login required)

View Item