An Ontology-Based Semantic Similarity Computation Model

摘要

The semantic similarity computation is very important in many applications that need to process text data. Because ontology can provide explicit concept specification, we establish an ontology-based semantic similarity computation model. This model quantifies the semantic similarity between concepts considering the semantic distance, concept level and the overlapping degree between sets of the hypernyms and hyponyms. The association between the semantic distance and concept level is also presented so that the number of parameters is reduced. The final experiment results demonstrate that this model is effective, and the accuracy is improved without any dependency on corpus.