rbo: Rank biased overlap (Webber et al., 2010)

Description

Evaluates the rank biased overlap (rbo) of two ranked lists based on formula based on (32) from "A Similarity Measure for Indefinite Rankings" (Webber et al.). Two ranked lists with high rbo are very similar, wheras low rbo indicates dissimilar lists. rbo ranges between 0 and 1. In this method the extrapolated version of rbo is implemented.

Usage

1
2
rbo(s, t, p, k = floor(max(length(s), length(t))/2), side = c("top",
  "bottom"), mid = NULL, uneven.lengths = TRUE)

Arguments

s

List 1

t

List 2

p

Weighting parameter in [0, 1]. High p implies strong emphasis on top ranked elements

k

Evaluation depth for extrapolation

side

Evaluate similarity between the top or the bottom of the ranked lists

mid

Set the mid point to for example only consider positive or negative scores

uneven.lengths

Indicator if lists have uneven lengths

Value

rank biased overlap (rbo)

Author(s)

Fabian Schmich

See Also

concordance

Examples

1
2
3
4
a <- rnorm(26)
b <- rnorm(26)
names(a) <- names(b) <- LETTERS
rbo(a, b, p = 0.95)

Questions? Problems? Suggestions? or email at ian@mutexlabs.com.

All documentation is copyright its authors; we didn't write any of that.