top.ranked.sets: Find k-top-ranked sets

In rbvs: Ranking-Based Variable Selection

Description

Finds k-top-ranked sets defined in Baranowski and Fryzlewicz (2015). This routine is used inside rbvs; it typically will be not called directly by the user.

Usage

top.ranked.sets(rankings, k.max, min.max.freq = 1, active = NULL)

Arguments

rankings	Matrix with rankings in each column.
k.max	Positive integer.
min.max.freq	Maximum frequency.
active	A vector with previously found active variables.

Details

Uses Portable qsort_r / qsort_s library ( Turner (2013)).

Value

List containing the following fields.

frequencies	Frequencies corresponding to the most frequent subsests at the top of the rankings.
subsets	The moost frequent subsets.

References

R. Baranowski, P. Fryzlewicz (2015), Ranking-Based Variable Selection, in submission (http://personal.lse.ac.uk/baranows/rbvs/rbvs.pdf)).
I. Turner (2013), Portable qsort_r / qsort_s, GitHub repository (https://github.com/noporpoise/sort_r).

rbvs documentation built on May 2, 2019, 7:31 a.m.