MPsuggestTopFeatures: Make a suggestion for interesting features in a dataset based...

In tkonopka/MultiPattern: Multi-Pattern Discovery Framework

Description

This function looks at the features that make up individual PCA components. It returns those features that contribute more than others to these componets (according to the weight of the feature compared to the median and IQR interval)

Usage

MPsuggestTopFeatures(x, ncomp = function(y) {     1 + sqrt(y) },
  iqrfactor = function(y) {     0.5 + log(y) })

Arguments

x	dataset or prcomp object
ncomp	function, integer, or numeric. If integer, number of principal components to consider. If number less than 1, fraction of components to consider. If function, a transformation on the number of features that returns an integer.
iqrfactor	numeric. threshold to consider a feature in a PC as worthy of selection

tkonopka/MultiPattern documentation built on May 31, 2019, 3:45 p.m.