Neural networks: algorithms, applications, and programming techniques
Neural networks: algorithms, applications, and programming techniques
Text Clustering to Help Knowledge Acquisition from Documents
EKAW '96 Proceedings of the 9th European Knowledge Acquisition Workshop on Advances in Knowledge Acquisition
Hi-index | 0.00 |
Today, users need pro-active information recovery mechanisms that can rapidly obtain the largest amount of relevant information possible. Several information recovery systems are available, but they still have semantic problems related to document treatment and retrieval. The proposed system uses statistical text-mining techniques to semantically associate documents to domains represented by ontologies, giving a similarity degree (relevance) between the document and the domain. A tool was implemented and tested using Jurisprudence documents from the Court of Justice of the State of Goiás. These documents are manually indexed by specialists under predefined categories. A comparison between the results obtained using the tool and the manual indexation was thus possible, permitting an analysis of the tool's efficiency using recall, precision and fall-out indexes.