fuzzyJaccard {statisfactory}

Fuzzy Jaccard index

Description

Calculates the fuzzy Jaccard index. The "normal" Jaccard index is given by sum(A intersect B) / sum(A union B), where A and B are sets. Typically, A and B are binary outcomes, but the fuzzy version can accommodate values in [0, 1] and/or binary outcomes. The computationally efficient and equivalent method is sum(pmin(A, B)) / (sum(A) + sum(B) - sum(pmin(A, B))). If A and B and both binary, the outcome is the same as the "plain" Jaccard index.

Usage

fuzzyJaccard(a, b)

Arguments

a, b

Vectors of binary and/or values in the range [0, 1]. The vectors must be of the same length.

Value

Numeric in the range [0, 1].

Examples

a <- c(0.3, 0, 0.9, 0.5)
b <- c(1, 1, 0, 0)
fuzzyJaccard(a, b)