DAAGxtras (version 0.6-9)

covsample: Sample of UCI Machine Learning Forest Cover Dataset

Description

Forest cover type is recorded, for every 50th observation taken from 581012 observations in the original dataset, together with a physical geographical variables that may account for the forest cover type.

Usage

data(covsample)

url

http://kdd.ics.uci.edu/databases/covertype/covertype.data.html

source

http://kdd.ics.uci.edu/databases/covertype/covertype.html

Details

For detailed information on the UCI dataset, see http://kdd.ics.uci.edu/databases/covertype/covertype.data.html

Variables V1 to V54 are physical geographical variables. Variable V55 is cover type, one of types 1 - 7. Note the omission of any information on geographical location. Distance through the data seems however to be, in part, a proxy for geographical location.

References

Blackard, Jock A. 1998. "Comparison of Neural Networks and Discriminant Analysis in Predicting Forest Cover Types." Ph.D. dissertation. Department of Forest Sciences. Colorado State University. Fort Collins, Colorado.

Examples

data(covsample)
options(digits=3)
tab.sample <- table(covsample$V55)
tab.sample/sum(tab.sample)
rm(covsample)
data(covtrain)
tab.train <- table(covtrain$V55)
tab.train/sum(tab.train)
rm(covtrain)
data(covtest)
tab.test <- table(covtest$V55)
tab.test/sum(tab.test)
rm(covtest)