An information retrieval (IR) system, implemented as a part of a content-driven hypertextual information retrieval (CoDHIR) project, is described. This work focuses on the use of semantic information that can be automatically acquired by applying natural language processing (NLP) techniques to texts. The information is represented using conceptual graphs. The problem of synonyms and homonyms is addressed in our system by using a model based on the interpretation of conceptual graphs extracted from texts. The detection of contextual roles of words allows an improvement in retrieval precision over traditional IR technologies. Ranking of documents, based on document relevance, is obtained by extending the vector space model into an oblique space and taking into account the relevance among different word couples.
Marega, R., Pazienza, M.t. (1994). CODHIR - AN INFORMATION-RETRIEVAL SYSTEM BASED ON SEMANTIC DOCUMENT REPRESENTATION. JOURNAL OF INFORMATION SCIENCE, 20(6), 399-412.
CODHIR - AN INFORMATION-RETRIEVAL SYSTEM BASED ON SEMANTIC DOCUMENT REPRESENTATION
PAZIENZA, MARIA TERESA
1994-01-01
Abstract
An information retrieval (IR) system, implemented as a part of a content-driven hypertextual information retrieval (CoDHIR) project, is described. This work focuses on the use of semantic information that can be automatically acquired by applying natural language processing (NLP) techniques to texts. The information is represented using conceptual graphs. The problem of synonyms and homonyms is addressed in our system by using a model based on the interpretation of conceptual graphs extracted from texts. The detection of contextual roles of words allows an improvement in retrieval precision over traditional IR technologies. Ranking of documents, based on document relevance, is obtained by extending the vector space model into an oblique space and taking into account the relevance among different word couples.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.