similarity {neighbr}

Calculate the similarity between two vectors of logicals.

Description

Calculate the similarity between two vectors of logicals.

Usage

similarity(x, y, measure)

Arguments

x, y

Logical or numeric vectors.

measure

Similarity measure ("simple_matching", "jaccard", or "tanimoto")

Details

Input vectors must consist of logical or numeric elements TRUE,FALSE or 0,1 (not factors). Similarity measures in this package are based on those defined in the PMML specification. Similarity ranges from 0 (no similarity) to 1 (identical).

For logical vectors x and y, we define the following:

a11 = number of times where x_i=1 and y_i=1
a10 = number of times where x_i=1 and y_i=0
a01 = number of times where x_i=0 and y_i=1
a00 = number of times where x_i=0 and y_i=0

Similarities are calculated using the following formulas:

Simple matching: (a11 + a00) / (a11 + a10 + a01 + a00)

Jaccard: (a11) / (a11 + a10 + a01)

Tanimoto: (a11 + a00) / (a11 + 2 * (a10 + a01) + a00)

Value

The similarity between x and y.

See Also

distance, PMML comparison measures

Examples

similarity(c(0,1,1),c(0,0,1),"simple_matching")
similarity(c(0,1,1),c(0,0,1),"jaccard")
similarity(as.logical(c(0,1,1)),as.logical(c(0,0,1)),"tanimoto")