From words to concepts in text mining

In the text mining process the interface between the natural language text and content reflecting terms is essential for the successful indexing process. The traditional interfaces such as the sauri and lexicons have limitations for web content processing. The recent lexical nets have promises; but their efficiency needs to be tested. The current work measures the use of word relations specified by lexical nets in a large test bed. The results correlate with similar recent past studies as not all specified word relations in the lexical nets are semantically rich in expressing the conceptual relations between given words. The results call for applying more heuristic approaches for text mining.