TESTING METHODS FOR AUTOMATIC DETECTION OF MORPHEME BOUNDARIES IN THE AZERBAIJANI LANGUAGE

Abstract
The paper is devoted to unsupervised learning and computer analysis of a natural language morphology, including automatic paradigm extraction and detection of morpheme boundaries. It describes in detail a finite-state approach implemented here. This approach has been tested on the material of the Azerbaijani language; the article discusses its results and compares them with those available in the Linguistica library on the same language.

Keywords: unsupervised learning of morphology, paradigm extraction, detection of morpheme boundaries