An Efficient Computational Method for Measuring Similarity between Two Conceptual Entities

摘要

Previous definitions of semantic similarity can be classified into two approaches. The node(information content)-based approach uses an entropy measure that is computed on the basis of child node population. The edge-based approach involves the use of the number of edges between two concepts within a hierarchical conceptual structure. The edge-based distance method is more intuitive, while the node-based information content approach is more theoretically sound. We consider a combined model that is derived from the edge-based notion with the addition of the information content. In this paper, we propose a method for computerized conceptual similarity calculation in WordNet space. The proposed method provides a degree of conceptual dissimilarity between two concepts. It gives a higher correlation value with a criterion based on human similarity judgment.