Discovering Word Meanings Based on Frequent Termsets

Word meaning ambiguity has always been an important problem in information retrieval and extraction, as well as, text mining (documents clustering and classification). Knowledge discovery tasks such as automatic ontology building and maintenance would also profit from simple and efficient methods for discovering word meanings. The paper presents a novel text mining approach to discovering word meanings. The offered measures of their context are expressed by means of frequent termsets. The presented methods have been implemented with efficient data mining techniques. The approach is domainand language-independent, although it requires applying part of speech tagger. The paper includes sample results obtained with the presented methods.