Using POS Annotations and EMdF Text Modeling for Semantic Web Search Export

@inproceedings{citeulike:2120411, abstract = {There is a large research effort for building a Semantic Web for complementing the current text-based web with machine understandable semantics. This paper introduces a new and complex algorithm for automatic Natural Language text exploration while exploring a corpus. The algorithm aims at searching and retrieving information from a corpus. At the same time the new algorithms can be used to store semantic based web repositories. The corpus is first explored by a morphologic and syntactic parser processing. An intermediate text corpus is structured according the Enhanced Monad dot Feature (EMdF) model. For the parsed text above an EMdF model is directly mapped and stored in an RDF database. Information retrieval on the initial text corpus is performed using queries on the RDF database. These queries resolve co- occurrences, verbs identification, and false positive elimination such that the results obtained are wellformed relations. The algorithms introduced are illustrated on examples of Semantic Web information retrieval on a medieval French corpus based on ” M\'{e}moires” of Philippe de Commynes.}, author = {Diaconescu, Rodica and Hirschb\"{u}hler, Paul and Ionescu, Dan and Martineau, France and Trifan, Mircea}, booktitle = {I2TS 2006}, citeulike-article-id = {2120411}, citeulike-linkout-0 = {http://www.inf.ufsc.br/i2ts2006/accepted.html}, citeulike-linkout-1 = {http://www.voies.uottawa.ca/Using_POS_Annot.pdf}, keywords = {article1, corpus, search, semanticweb}, posted-at = {2007-12-14 20:42:53}, priority = {0}, title = {Using POS Annotations and EMdF Text Modeling for Semantic Web Search}, url = {http://www.inf.ufsc.br/i2ts2006/accepted.html} }

TY - CONF ID - citeulike:2120411 L3 - citeulike-article-id:2120411 N2 - There is a large research effort for building a Semantic Web for complementing the current text-based web with machine understandable semantics. This paper introduces a new and complex algorithm for automatic Natural Language text exploration while exploring a corpus. The algorithm aims at searching and retrieving information from a corpus. At the same time the new algorithms can be used to store semantic based web repositories. The corpus is first explored by a morphologic and syntactic parser processing. An intermediate text corpus is structured according the Enhanced Monad dot Feature (EMdF) model. For the parsed text above an EMdF model is directly mapped and stored in an RDF database. Information retrieval on the initial text corpus is performed using queries on the RDF database. These queries resolve co- occurrences, verbs identification, and false positive elimination such that the results obtained are wellformed relations. The algorithms introduced are illustrated on examples of Semantic Web information retrieval on a medieval French corpus based on “Mémoires” of Philippe de Commynes. TI - Using POS Annotations and EMdF Text Modeling for Semantic Web Search T2 - I2TS 2006 KW - article1 KW - corpus KW - search KW - semanticweb AU - Diaconescu, Rodica AU - Hirschbühler, Paul AU - Ionescu, Dan AU - Martineau, France AU - Trifan, Mircea UR - http://www.inf.ufsc.br/i2ts2006/accepted.html ER -