SCM

SCM Repository

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

View of /pkg/man/UNIMANshort.Rd

Parent Directory Parent Directory | Revision Log Revision Log


Revision 55 - (download) (as text) (annotate)
Wed Mar 2 14:48:07 2016 UTC (2 years, 11 months ago) by variani
File size: 2764 byte(s)
new Rd produced by roxygen2
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/datasets.R
\docType{data}
\name{UNIMANshort}
\alias{UNIMANshort}
\title{Dataset UNIMANshort.}
\description{
Short-term UNIMAN datasets of 200 samples from 17 polymeric sensors. 
The datasets contains two matricies:
\tabular{rl}{
  \code{C} \tab The concentration matrix of 200 rows and 3 columns encodes 
    the concentration profile for three gases, ammonia, propanoic acid and n-buthanol.
    The concentration units are given in the percentage volume (\% vol.).
    Ammonia has three concentration levels 0.01, 0.02 and 0.05,  propanoic acid - three levels 0.01, 0.02 and 0.05,
    and n-buthanol - two levels 0.1 and 1. \cr
  \code{dat} \tab The data matrix of 200 rows and 17 columns cotains the steady-state signals of 17 sensors
    in response to the concentration profle \code{C}. \cr
}
}
\details{
The reference dataset has been measured at The University of
Manchester (UNIMAN). Three analytes ammonia, propanoic
acid and n-buthanol, at different concentration levels, were
measured for 10 months with an array of seventeen conducting polymer sensors.

In modeling of the array we make the distinction between
short-term and long-term reference data. Two hundred samples
from the first 6 days are used to characterize the array assuming
the absence of drift. The long-term reference data (not published within the package) 
counts for the complete number of samples from 10 months,
these data were used to model the sensor noise and drift,
see \code{\link{UNIMANsnoise}} and \code{\link{UNIMANdnoise}} for more details.

A pre-processing procedure on outliers removal was applied to the
reference data. The standard method based on the
squared Mahalanobis distance was used with quantile equal to
\code{0.975}\%.
}
\examples{

data("UNIMANshort", package="chemosensors")

str(UNIMANshort)

C <- UNIMANshort$C
dat <- UNIMANshort$dat

# plot sensors in affinity space of gases
#plotAffinitySpace(conc=C, sdata=dat, gases=c(2, 1))
#plotAffinitySpace(conc=C, sdata=dat, gases=c(2, 3))
#plotAffinitySpace(conc=C, sdata=dat, gases=c(3, 1))

# make standar PCA (package 'pls') to see:
# - multi-variate class distribution (scoreplot)
# - low-dimensionality of data (variance)
# - contribution of 17 sensors in terms of linear modeling (loadings)
mod <- prcomp(dat, center=TRUE, scale=TRUE)

col <- ccol(C)
scoreplot(mod, col=col, main="PCA: Scoreplot")

barplot(mod$sdev, main="PCA: Sd. Deviation ~ PCs")

loadings <- mod$rotation
col <- grey.colors(3, start=0.3, end=0.9)
matplot(loadings[, 1:3], t='l', col=col, lwd=2, lty=1,
  xlab="Sensor", ylab="sdev", main="PCA: Loadings PCs 1-3 ~ Sensors")
}
\seealso{
\code{\link{SensorModel}}, \code{\link{SensorModel}}
}
\keyword{data}


R-Forge@R-project.org
ViewVC Help
Powered by ViewVC 1.0.0  
Thanks to:
Vienna University of Economics and Business University of Wisconsin - Madison Powered By FusionForge