An ontology based approach to measuring the semantic similarity between information objects in personal information collections

This paper introduces a semantic approach to personal information management, which employs natural language processing, ontologies and a vector space model to measure the semantic similarity between information objects in personal information collections. The approach involves natural language processing, named entity recognition, and information object integration. In particular, natural language processing is used to detect meaningful and semantically distinguishable information objects within collections of personal information. Then, the named entities are extracted from these information objects and their features (such as weight and category) are used to measure the semantic similarity between them. Further research includes using the semantic similarity measure developed to index and retrieve information objects in a semantic based system for personal information management