randomforest_res: Importance of significant features.

In umerijaz/microbiomeSeq: Microbial community analysis in an environmental context.

Description

This function uses Breiman's random forest algorithm that can be used in unsupervised mode for assessing proximities among data points. Random forest classifier is used to determine the importance of differentially expressed features/taxa to the microbial community. The measure used in this case is Mean Descrease in Accuracy aso know as permutation importance which is reported for each of the features. This is obtained by removing the relationship of a feature and measuring increase in error. Consequently, the feature with high mean decrease in accuracy is considered most important.

Usage

randomforest_res(data, groups)

Arguments

data	(Required). A data.frame of features to be assigned importance.
groups	(Required). Vector corresponding to groups/conditions or source of variation in data.

Value

Returns a data.frame of importance measure (mean decrease in accuracy) and ranks.

Author(s)

Alfred Ssekagiri [e-mail], Umer Zeeshan Ijaz [e-mail]

References

http://userweb.eng.gla.ac.uk/umer.ijaz/, Umer Ijaz, 2015

umerijaz/microbiomeSeq documentation built on May 30, 2019, 3:13 p.m.