jaccard_coef: Jaccard coefficient

Description

Calculate the Jaccard (similarity) coefficient between words.

Usage

jaccard_coef(x, ...)
"jaccard_coef"(x, max.size = 1000, dist = FALSE)
"jaccard_coef"(x, max.size = 1000, stopwds = unique(c(tm::stopwords(), letters)), ignore.case = TRUE, dist = FALSE)

Arguments

Character vector with the phrases (tweets) to be analyzed

max.size

Max number of words to analyze

dist

When true computes one minus Jaccard coef

stopwds

Character vector of stopwords

ignore.case

When true converts all to lower

Value

A list including a dgCMatrix matrix

Methods (by class)

list: Method Processes a list of character vectors such as the one obtained from tw_extract
character: Computes the coef from a vector of characters (splits the text)

Details

The Jaccard index is used as a measure of similarity between two elements. In particular for a given pair of elements $x,y$ it is calculated as $$J(S,T) = \frac{|S\cap T|}{|S\cup T|}$$ Where $S$ is the set of groups where $x$ is present and $T$ is the set of groups where $y$. The resulting value is defined between 0 and 1, where 0 corresponds to no similarity at all (the elements don't have a group in common) and 1 represents perfect similarity (both elements are present in the same groups).

References

Conover, M., Ratkiewicz, J., & Francisco, M. (2011). "Political polarization on twitter". Icwsm, 133(26), 89–96. http://doi.org/10.1021/ja202932e