|
a |
|
b/OmicsFold/man/diablo.selection.stability.Rd |
|
|
1 |
% Generated by roxygen2: do not edit by hand |
|
|
2 |
% Please edit documentation in R/feature_loadings.R |
|
|
3 |
\name{diablo.selection.stability} |
|
|
4 |
\alias{diablo.selection.stability} |
|
|
5 |
\title{Extract DIABLO feature selection stability} |
|
|
6 |
\usage{ |
|
|
7 |
diablo.selection.stability(diablo.perf, comp, block) |
|
|
8 |
} |
|
|
9 |
\arguments{ |
|
|
10 |
\item{diablo.perf}{Performance evaluation of the DIABLO model generated by |
|
|
11 |
mixOmics perf.} |
|
|
12 |
|
|
|
13 |
\item{comp}{Component number for which to retrieve feature selection |
|
|
14 |
stabilities.} |
|
|
15 |
|
|
|
16 |
\item{block}{Block number for which to retrieve feature selection |
|
|
17 |
stabilities.} |
|
|
18 |
} |
|
|
19 |
\value{ |
|
|
20 |
A data frame containing the the features selected for the specificed |
|
|
21 |
component and their relative stability, as a proportion of trained models |
|
|
22 |
0-1. |
|
|
23 |
} |
|
|
24 |
\description{ |
|
|
25 |
Extract feature selection stabilities for a given component from a |
|
|
26 |
performance validated mixOmics multi-omics DIABLO model. Feature stabilities |
|
|
27 |
are an important indicator of the confidence that a selected feature is |
|
|
28 |
predictive for the outcome classes in the model, and hence (in combination |
|
|
29 |
with the loading weight) is likely to be of biological significance. Features |
|
|
30 |
are retrieved for a single specified block and component. |
|
|
31 |
} |
|
|
32 |
\examples{ |
|
|
33 |
\dontrun{ |
|
|
34 |
diablo.selection.stability(perf.diablo.analysis, comp=1, block=1) |
|
|
35 |
} |
|
|
36 |
} |