Word Sense Disambiguation (WSD) for Indonesian Homograph Word Meaning Determination by LESK Algorithm Application

摘要

Indonesian has several words which are commonly known as ambiguous words, confusing the meaning of a sentence or a statement to be less understood or even not delivered. It is different to a human perception which has linguistic ability to determine the meaning of ambiguity or more than a word meaning. Word Sense Disambiguation is one of a topic from natural language processing (NLP) which deals with ambiguity handling. Word Sense Disambiguation is a linguistic computational process which aims to identify the proper meaning of words based on the context. This current study is designed as a system to handle the ambiguous words. It is conducted by looking up and defining the meaning of ambiguous words by using LESK algorithm. The test is performed towards the functionality from a system in which the result system test is in line with test data from KBBI. The result presents accuracy level of 78.6% for one of an ambiguous word and 62.5 % for two of ambiguous words in determining the meaning appropriately.