A simple and effective Concept Vector for WordNet semantic measurement

We first propose a basic Concept Vector, and then we define the Relevancy Node of a given node in a tree, and improved the basic vector through paths between the compared nodes and their Relevancy Nodes in WordNet hierarchy. The similarity of two concepts is obtained by computing cosine similarity between their concept vectors. The vector contains structure information inherent in WordNet. We show that this measure compares favorably to other measures. The method is flexible in that it can make comparisons between any two concepts in a WordNet like hierarchy without relying on additional dictionary or corpus information.