A Weight-Based Concept Similarity Algorithm

In order to improve the accuracy of concept similarity and complete the semantic information of concept similarity, a weight-based concept similarity algorithm is proposed in this paper. According weight distribution without considering the structural relationship between concepts that the accuracy of concept similarity is not high, to import the property relationship and inheritance relationship, and the impact of the concept's depth to similarity degree. We defined a weight distribution approach to measure the semantic distance of two concepts, according to which we present a math formula to compute the similarity degree. Experimental evaluation demonstrates that the proposed algorithm outperforms traditional similarity measures and accords with field experts' anticipated results.