A Hybrid method for Evaluating Semantic Similarity of Words Based on Ontology

Semantic Similarity is considered as an important component of many applications such as Natural Language Processing, Information Retrieval, Text Clustering and etc. This paper presents a new method for evaluating the semantic similarity of words based on ontology. Ontology development in many field and its structural features has led to be used as an effective knowledge-base in proposed method. In this paper, we introduce a hybrid method that computes similarity considering both structural features (shortest path between two nodes, neighbors, hyponym and etc.) and glosses of entity. Finally, we use WordNet and evaluate this work related to human similarity scores. The results are given to demonstrate the effectiveness of our approach against related work.