SimCT: A measure of semantic similarity adapted to hierarchies of concepts

The Calculating of the similarity between data is a key problem in several disciplines such as machine learning, information retrieval (IR) and data analysis. In some areas such as social resilience, the similarity measures can be used to find the similarities between traumatized individuals or resilience's dimensions. In this paper, we propose a measure of semantic similarity used in many applications including clustering and information retrieval. It relies on a knowledge base represented as a hierarchy of concepts (ontology, graph, taxonomy). Its uniqueness with respect to previous proposals is the difference between the indices of similarity that it establishes between brothers concepts located at the same hierarchical level and having the same direct ancestor. In addition, our semantic similarity measure provides better modularity in clustering compared with Wu and Palmer's similarity measure and Proxygenea 3.