Classification Method for Learning Morpheme Analysis

Abstract

The morpheme analysis module is an important component in natural language processing engines. The parser modules are usually based on rule systems created by human experts. In the paper, a novel approach is tested for implementation of the morpheme analyzer module. The proposed structure is based on the theory of formal concept analysis. The word inflection can be considered as a classification problem, where the class label denotes the corresponding transformation rule. The main benefit of the proposed method is the efficient generalization feature. The proposed morpheme analyzer module was implemented in a prototype question generation application.