Sem@ntica: A system for semantic extraction and logical querying of text corpora

Abstract

Sem@ntica is a system for extracting the information contained in collections of documents into a knowledge base. It combines high quality conventional named entity analysis with an ontology class labeling capability for open class words. The ontology comprises an upper ontology and one or more domain ontologies. The system has tools for rapidly designing the ontology and mapping segments of Word Net on to ontology classes. The extracted knowledge base can be queried directly using the KM language and tools are provided for rapid construction of dynamic web page reports and for portal design. The system is open and its capabilities can be controlled and viewed via an extended suite of dynamic web page visualizations.Its component capabilities are available via TCP/IP remote procedure calls and via SOAP web services. Sem@ntica is designed to provide the front-end processing for national and commercial intelligence processing; it has a wide field of potential application.Index Terms-information extraction, artificial intelligence, named entity analysis, knowledge base, ontology.