Semantic similarity analysis—A computer-based study of meaning in noun phrases

The purpose of this paper is to present a method by which a computerized data base which contains phrases of natural language information can be created and interrogated in the semantic domain for multiple uses and users. The method takes into account both the individual user’s needs and concepts of similarity in meaning of textual material represented by the stored data. The method is applicable to arbitrary bodies of natural language text which occur in the syntactic format of the noun phrase. The method was developed and tested, however, using noun phrases of medical text since its capabilities are directly applicable to needs in that area.