Book: Machine-Aided Linguistic Discovery
Chapter: Detecting Significant Similarities: RECLASS and the Problem of Genetic Language Classification
Blurb:
Given a set of languages described in terms of feature values or wordlists, as well as the current genetic classification of these languages, the RECLASS program, which will be the object of discussion in this chapter, finds all pairs of languages, such that the languages in each pair both exhibit statistically significant similarities and belong to different language families.