Learning rules for morphological analysis and synthesis of Macedonian nouns, adjectives and verbs

Abstract

This paper presents a machine learning approach to morphological analysis and synthesis of Macedonian nouns, adjectives and verbs. An inductive logic programming (ILP) system, Clog, was used to learn the inflectional paradigms of Macedonian words. Clog learns first-order decision lists, i.e., ordered sets of rules. Training and testing of the rules was performed on the words originating from Orwell’s “1984”. High accuracies (over 90%) were achieved, which are encouraging both for further research in annotation of Macedonian language resources as well as practical use (eg. in web search engines).