Semantic Similarity Calculation Based on Sememe Set

The calculation of semantic similarity is a key point in Chinese information processing. This paper presents a new method for calculating semantic similarity between two words. Different from previous methods, this paper focuses on the perspective of connotations, trying to highlight the essential attributes of words by which we can get the similarity value in words´ conceptual level. Firstly, we get definitions from machine-readable dictionaries. Every definition will be translated into an interpretation vector. Then we use sememe, which is the least significant unit of concept in HowNet, as the unit of definition in the iterative process. Evaluations show that the method is effective and achieves good results.