Analysis of Concept Similarity Methods Applied to an LSH Function

In literature, there are several methods to measure similarity between concepts in structures like simple ontologies, concept hierarchies, taxonomies, etc. These measures are used to search for similar concepts. In the Semantic Web, such structures are commonly used to classify data which opens the possibility of reasoning upon them and helps in conceptual searches. Besides that, the Locality Sensitive Hash (LSH) functions are used to store similar data close to each other in an index space. Each family of LSH functions is tied to a specific similarity function. In this paper we propose a method for combining the idea of conceptual similarity with LSH functions. This method permits the data classified as similar concepts be indexed close to each other respecting some metric. The main idea is to facilitate the conceptual searching for data semantically classified. This paper evaluates several methods of measuring the similarity between concepts in a simple ontology and discusses how they can be applied to an LSH function.