Reducing a spatial database to its effective dimensionality for logistic - regression analysis of incidence of livestock disease
Large databases with multiple variables, selected because they are available and might provide an insight into establishing causal relationships, are often difficult to analyse and interpret because of multicollinearity. The objective of this study was to reduce the dimensionality of a multivariable...
| Main Authors: | , , |
|---|---|
| Format: | Journal Article |
| Language: | Inglés |
| Published: |
Elsevier
1997
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10568/29499 |
| Summary: | Large databases with multiple variables, selected because they are available and might provide an insight into establishing causal relationships, are often difficult to analyse and interpret because of multicollinearity. The objective of this study was to reduce the dimensionality of a multivariable spatial database of Zimbabwe, containing many environmental variables that were collected to predict the distribution of outbreaks of theileriosis (the tick-borne infection of cattle caused by Theileria parva and transmitted by the brown ear tick). Principal-component analysis and varimax rotation of the principal components were first used to select a reduced number of variables. The logistic-regression model was evaluated by appropriate goodness-of-fit-tests. |
|---|