Construction of Myanmar WordNet lexical database

Abstract

A lexical resource is the central repository of data for all language processing applications. It contains information for human consumption as well as computer programs. A significant increase in the use of lexical databases has led WordNet to become one of the most widely used lexical information source for Natural Languages Processing (NLP) applications. This paper presents a methodology of constructing Myanmar WordNet lexical database from WordNet and Myanmar<->English Machine Readable Dictionaries (MRDs). This system consists of three phases which are (i) data extraction phase (ii) link analyzing phase and (iii) construction phase. In the data extraction phase, the lexical information is extracted from available resources. Then sets of candidate links between Myanmar words and synsets are derived from lexical information and that are verified for the lexical gap between two languages in the link analyzing phase. Finally, a set of the verified candidate link is used to automatically construct the Myanmar WordNet. The proposed Myanmar WordNet has coverage of 173853 links, 38142 synsets and 25865 words.