SCM Repository

[chemosensors] View of /pkg/man/UNIMANdnoise.Rd
ViewVC logotype

View of /pkg/man/UNIMANdnoise.Rd

Parent Directory Parent Directory | Revision Log Revision Log

Revision 55 - (download) (as text) (annotate)
Wed Mar 2 14:48:07 2016 UTC (3 years, 1 month ago) by variani
File size: 1836 byte(s)
new Rd produced by roxygen2
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/datasets.R
\title{Dataset UNIMANdnoise.}
The dataset contains statistics on a multi-variate drift sub-space 
for the long-term UNMIAN dataset of 1000 samples over 8 classes.
The subspace was evaluated via common principal component analysis (power algorithm).
Only three classes (analytes at maximum concentration) are used in the computation for more accurate estimation.
The importance of the drift components is computed as a projected variance 
of sensor array data of the given three classes onto the components.

The datasets contains four variables:
  \code{dspace} \tab A square \code{17 x 17} matrix of the drift subspace. Columns are drift component vectors. \cr
  \code{ndvar} \tab  A vector of length \code{17} with the importance of the components.\cr
  \code{centered} \tab  Boolean indicating whether the UNIMAN data were centered before the computation. \cr
  \code{scaled} \tab  Boolean indicating whether the UNIMAN data were scaled before the computation. \cr



# 17 sensor loadings, 17 drift components

# importance of 17 drift components

barplot(UNIMANdnoise$ndvar, names.arg=1:length(UNIMANdnoise$ndvar), 
  main="Importance of drift components")
# comp. 1 dominates, 
# comp. 1-3 dominate over the rest

loadings <- UNIMANdnoise$dspace
col <- grey.colors(3, start=0.3, end=0.9)
matplot(loadings[, 1:3], t='l', col=col, lwd=2, lty=1,
  xlab="Sensors", ylab="Loadings", 
  main="Loadings of sensors on first 3 drift components")
# sensors 7, 8, 9, 17 seem to be less drifty  
ViewVC Help
Powered by ViewVC 1.0.0  
Thanks to:
Vienna University of Economics and Business University of Wisconsin - Madison Powered By FusionForge