Abstract The paper is dedicated to comparative analysis of phyogenetic algorithms used for linguistics tasks. At present there are a lot of phylogenetic algorithms; however, there is no unanimous opinion on which of them should be used. The paper suggests the model of language evolution trees and introduces a parameter to characterize the topology of trees. The comparison of the main algorithms is made on the trees of various topology. The paper displays that the UPGMA algorithm gives better results on the trees close to balanced ones. It provides the explanation for a number of contradictive results, described in published works. The problem of the input data choice and the relation between results and the number and type of parameters is under consideration. The results obtained are also ambiguous. Typological databases “Jazyki mira” and WALS as well as the method of computer modeling are used in the paper.

Keywords: language evolution, phylogenetic algorithms

Link:

COMPARATIVE ANALYSIS OF PHYLOGENIC ALGORITHMS

Valery Solovyev, Renat Faskhutdinov

http://www.foibg.com/ijitk/ijitk-vol04/ijitk04-3-p03.pdf