Ontology generation for large email collections

Abstract

This paper presents a new approach to identifying concepts expressed in a collection of email messages, and organizing them into an ontology or taxonomy for browsing. It incor-porates techniques from text mining, information retrieval, natural language processing and machine learning to gener-ate a concept ontology. Nominal N-gram mining is used to identify candidate concepts. Wordnet and surface text pat-tern matching are used to identify relationships among the concepts. A supervised clustering algorithm is then used to further cluster the concepts. The experiments show that the approach is effective.