clusterByQuality {AneuFinder} | R Documentation |
This function uses the mclust package to cluster the input samples based on various quality measures.
clusterByQuality(hmms, G = 1:9, itmax = c(100, 100), measures = c("spikiness", "entropy", "num.segments", "bhattacharyya", "complexity", "sos"), orderBy = "spikiness", reverseOrder = FALSE)
hmms |
A list of |
G |
An integer vector specifying the number of clusters that are compared. See |
itmax |
The maximum number of outer and inner iterations for the |
measures |
The quality measures that are used for the clustering. Supported is any combination of |
orderBy |
The quality measure to order the clusters by. Default is |
reverseOrder |
Logical indicating whether the ordering by |
Please see getQC
for a brief description of the quality measures.
A list
with the classification, parameters and the Mclust
fit.
Aaron Taudt
## Get a list of HMMs folder <- system.file("extdata", "primary-lung", "hmms", package="AneuFinderData") files <- list.files(folder, full.names=TRUE) cl <- clusterByQuality(files) ## Plot the clustering and print the parameters plot(cl$Mclust, what='classification') print(cl$parameters) ## Select files from the best 2 clusters for further processing best.files <- unlist(cl$classification[1:2])