getFeatures: Features list extraction

In komorowskilab/R.ROSETTA: R.ROSETTA: an interpretable machine learning framework

Description

Extracts the features from the rules. Sorts the features according to their frequency (percentage). Creates a list of the features, frequencies and discretization levels for each decision class.

Usage

getFeatures(rules, filter = F, filterType = "support", thr = 10)

Arguments

rules	A data frame of rule table in a form of main rosetta output.
filter	Logical. Enable filtration or not. Default is F.
filterType	A character. Choose between support, accuracy or pvalue. The default is 10.
thr	A numeric value of rule threshold. The default is 10.

Value

list	Lists of the features, frequencies and discretization levels for each class. Each list is sorted be the frequency.

Author(s)

Mateusz Garbulowski

Examples

library(R.ROSETTA)
set.seed(1)

out <- rosetta(autcon)
rules <- out$main

# default settings
features <- getFeatures(rules)

# filter by p-value
features <- getFeatures(rules, filter = T, filterType = "pvalue", thr = 0.05)

komorowskilab/R.ROSETTA documentation built on April 5, 2021, 9:41 a.m.