Variable selection in generalized linear models with canonical link functions

Jin, Man; Fang, Yixin; Lincheng, Zhao

doi:10.1016/j.spl.2004.11.021

Cited by 5 publications

(1 citation statement)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…GLM is a mathematical extension of linear models that do not force data into unnatural scales, and thereby allow for nonlinearity and non-constant variance structures in the data (Jin et al, 2005;McCullagh and Nelder, 1989). They are based on an assumed relationship (link function) between the mean of the response variable and the linear combination of the explanatory variables.…”

Section: Generalized Linear Models (Glm)mentioning

confidence: 99%

Arsenic-Safe Aquifers in Coastal Bangladesh: An Investigation With Ordinary Kriging Estimation

Hassan

Ahamed

2017

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

ABSTRACT:Spatial point pattern is one of the most suitable methods for analysing groundwater arsenic concentrations. Groundwater arsenic poisoning in Bangladesh has been one of the biggest environmental health disasters in recent times. About 85 million people are exposed to arsenic more than 50μg/L in drinking water. The paper seeks to identify the existing suitable aquifers for arsenic-safe drinking water along with "spatial arsenic discontinuity" using GIS-based spatial geostatistical analysis in a small study site (12.69 km 2 ) in the coastal belt of southwest Bangladesh (Dhopakhali union of Bagerhat district). The relevant spatial data were collected with Geographical Positioning Systems (GPS), arsenic data with field testing kits, tubewell attributes with observation and questionnaire survey. Geostatistics with kriging methods can design water quality monitoring in different aquifers with hydrochemical evaluation by spatial mapping. The paper presents the interpolation of the regional estimates of arsenic data for spatial discontinuity mapping with Ordinary Kriging (OK) method that overcomes the areal bias problem for administrative boundary. This paper also demonstrates the suitability of isopleth maps that is easier to read than choropleth maps. The OK method investigated that around 80 percent of the study site are contaminated following the Bangladesh Drinking Water Standards (BDWS) of 50μg/L. The study identified a very few scattered "pockets" of arsenic-safe zone at the shallow aquifer.

show abstract

Section: Generalized Linear Models (Glm)mentioning

confidence: 99%