outlierclassifier: Detect and classify compositional outliers.

Description

Detects outliers and classifies them according to different possible explanations.

Usage

OutlierClassifier1(X,...)
# S3 method for acomp
OutlierClassifier1(X,...,alpha=0.05,
           type=c("best","all","type","outlier","grade"),goodOnly=NULL,
           corrected=TRUE,RedCorrected=FALSE,robust=TRUE)

Arguments

the dataset as an acomp object

…

further arguments to MahalanobisDist/gsi.mahOutlier

alpha

The confidence level for identifying outliers.

type

What type of classification should be used: best: Which single component would best explain the outlier. all: Give a binary coding specifying all components, which could explain the outlier. type: Is it a a normal observation "ok", a single componentn outlier "1" or can it not be explained by a single wrong component "?". outlier: All outliers are marked as "outlier", others are marked as "ok". "grade": Proven Outliers are marked as "outlier"s, suspected outliers, detected without correction of the p-value are reported as "extreme", the rest is reported as "ok".

goodOnly

an integer vector. Only the specified index of the dataset should be used for estimation of the outlier criteria. This parameter if only a small portion of the dataset is reliable.

corrected

logical. Literatur often proposed to compare the Mahalanobis distances with Chisq-Approximations of there distributions. However this does not correct for multiple testing. If corrected is true a correction for multiple testing is used. In any case we do not use the chisq-approximation, but a simulation based procedure to compute confidence bounds.

RedCorrected

logical. If an outlier is detected we can try to find out wether a single component would be sufficient to drop the outlier under the outlier detection limit. Since in this second case we only check a few outliers no second correction step applies as long as the number of outliers is not very high.

robust

A robustness description as define in var.acomp

Value

A factor classifying the observations in the dataset as "ok" or some type of outlier.

Details

See outliersInCompositions for a comprehensive introduction into the outlier treatment in compositions.

See ClusterFinder1 for an alternative method to classify observations in the context of outliers.

Examples

Run this code

# NOT RUN {
tmp<-set.seed(1400)
A <- matrix(c(0.1,0.2,0.3,0.1),nrow=2)
Mvar <- 0.1*ilrvar2clr(A%*%t(A))
Mcenter <- acomp(c(1,2,1))
data(SimulatedAmounts)
datas <- list(data1=sa.outliers1,data2=sa.outliers2,data3=sa.outliers3,
              data4=sa.outliers4,data5=sa.outliers5,data6=sa.outliers6)
opar<-par(mfrow=c(2,3),pch=19,mar=c(3,2,2,1))  
tmp<-mapply(function(x,y) {
outlierplot(x,type="scatter",class.type="grade");
  title(y)
},datas,names(datas))


par(mfrow=c(2,3),pch=19,mar=c(3,2,2,1))  
tmp<-mapply(function(x,y) {
  myCls2 <- OutlierClassifier1(x,alpha=0.05,type="all",corrected=TRUE)
  outlierplot(x,type="scatter",classifier=OutlierClassifier1,class.type="best",
  Legend=legend(1,1,levels(myCls),xjust=1,col=colcode,pch=pchcode),
  pch=as.numeric(myCls2));
  legend(0,1,legend=levels(myCls2),pch=1:length(levels(myCls2)))
  title(y)
},datas,names(datas))

par(mfrow=c(2,3),pch=19,mar=c(3,2,2,1))  
for( i in 1:length(datas) ) 
  outlierplot(datas[[i]],type="ecdf",main=names(datas)[i])
par(mfrow=c(2,3),pch=19,mar=c(3,2,2,1))  
for( i in 1:length(datas) ) 
  outlierplot(datas[[i]],type="portion",main=names(datas)[i])
par(mfrow=c(2,3),pch=19,mar=c(3,2,2,1))  
for( i in 1:length(datas) ) 
  outlierplot(datas[[i]],type="nout",main=names(datas)[i])
par(opar)
# }