OmicsMarkeR (version 1.4.2)

aggregation: Feature Aggregation

Description

Compiles matrix of ranked features via user defined 'metric'

Usage

aggregation(efs, metric, f = NULL)

Arguments

efs

A matrix of selected features

metric

string indicating the type of aggregation. Avialable options are "CLA" (Complete Linear), "EM" (Ensemble Mean), "ES" (Ensemble Stability), and "EE" (Ensemble Exponential)

f

The number of features desired. Default f = NULL

Value

agg

Aggregated list of features

References

Abeel T., Helleputte T., Van de Peer Y., Dupont P., Saeys Y. (2010) Robust biomarker identification for cancer diagnosis with ensemble feature selection methods. Bioinformatics 26(3) 392-398.

Meinshausen N., Buhlmann P. (2010) Stability selection. J.R. Statist. Soc. B. 72(4) 417-473.

Haury A., Gestraud P., Vert J. (2011) The Influence of Features Selection Methods on Accuracy, Stability, and Interpretability of Molecular Signatures. PLoS ONE 6(12) e28210. doi: 10.1371/journal.pone.0028210.

See Also

CLA, ES, EM, EE

Examples

# test data
ranks <- replicate(5, sample(seq(50), 50))
row.names(ranks) <- paste0("V", seq(50))

aggregation(ranks, "CLA")