A New Measure of Word Semantic Similarity Based on WordNet Hierarchy and DAG Theory

The problem of measuring the semantic similarity between words has been considered a fundamental operation in the field of computational lexical semantics, but the accuracy of existing computational methods is not very close to what humans would perceive. This paper presents a new approach to measure the semantic similarity between words in the hierarchy of WordNet. Our approach considers not only the semantic distance between two words but also the feature information of the DAG (Directed Acyclic Graph). A common data set of word pairs is used to evaluate the proposed approach: we first calculate the semantic similarities of 30 word pairs, then the correlation coefficient between human judgement and six computational measures are calculated, the experiment shows our approach is better than other existing computational models.