hclustWidget: shiny-oriented GUI for cluster or classifier exploration

In lgatto/MLInterfaces: Uniform interfaces to R machine learning procedures for data in Bioconductor containers

Description

shiny-oriented GUI for cluster or classifier exploration

Usage

hclustWidget(mat, featureName = "feature", title =
                 paste0("hclustWidget for ", deparse(substitute(mat))),
                 minfeats = 2, auxdf = NULL)

mlearnWidget(eset, infmla)

Arguments

mat	matrix with feature vectors in rows
featureName	name to be used for control that asks for number of features to use
title	widget title
minfeats	lower bound on number of features to use
auxdf	data.frame with number of rows equal to nrow(mat), with metadata to be displayed in hovering tooltip
eset	instance of ExpressionSet-class
infmla	instance of formula, with dependent variable values obtained in pData(eset), and independent variable specified as "." or as properly munged elements of featureNames(eset).

Details

Experimental tool to illustrate impacts of choice of distance, agglomeration method, etc.

Value

a shinyApp result that will display in active browser

Note

mlearnWidget will attempt to nicely produce a variable importance plot using randomForestI. This means that the annotation package for probe identifiers should be loaded or an error will be thrown.

Author(s)

VJ Carey <[e-mail]>

Examples

# should run with example(hclustWidget, ask=FALSE)
if (interactive()) {
 library(shiny)
 library(MASS)
 data(crabs)
 cr = data.matrix(crabs[,-c(1:3)])
 au = crabs[,1:3]
 show(hclustWidget(cr, auxdf=au))
## must use stop widget button to proceed
  library(ALL)
  library(hgu95av2.db)
  data(ALL)
  show(mlearnWidget(ALL[1:500,], mol.biol~.))
 }

lgatto/MLInterfaces documentation built on May 21, 2019, 5:12 a.m.