Learning Multilingual Morphology with CLOG
ILP '98 Proceedings of the 8th International Workshop on Inductive Logic Programming
The Oxford Handbook of Computational Linguistics (Oxford Handbooks in Linguistics S.)
The Oxford Handbook of Computational Linguistics (Oxford Handbooks in Linguistics S.)
Induction of first-order decision lists: results on learning the past tense of English verbs
Journal of Artificial Intelligence Research
| Hi-index | 0.00 |
One of the most explored fields of NLP is morphology. It is important because language is productive: in any given text we will encounter text words an word forms that we haven't seen before and that are not in a precompiled dictionary. The core task of computational morphology is to take a word as input and produce a morphonological analysis for it. There are a lot of approaches in formalizing and automatically finding rules that implement the morphology of a language. One approach is through decision lists. Decision lists can be employed as a representation language for a wide range of tasks. Clog is a system that is based on this logic and that has been primarily developed with natural language applications in mind. So far, Clog has been successfully employed for morphology learning tasks. The research presented in this paper is about machine learning of the rules of nouns morphology in Albanian. Starting point of the research presented here are 7 lists, each consisting of 6000 pairs of feminine nouns in Albanian. The pairs in the lists hold the noun's base form and inflected form. As base word is considered the definite singular of nominative.