A Web-based Tagger for Named Entities Detection

Abstract

Information Retrieval constitutes an important task for extracting meaningful data in huge amounts of texts since the majority of information in the world exist electronically, and in unstructured and plain text form. Most datamining applications assimilate only structured information, this means before any predictive model can be applied the data must be prepared in a special way. This paper presents a web-based tagger system for named entities detection in plain text to help users tagging texts resources. The advantage of using this tool is the possibility of tagging any plain text file and with any type of class for the entities, allowing to treat any domain and thus obtaining a wider recognition of named entities. In addition, a small application shows the facility in the training of NER classifier models as an advantage of corpora constructed with the web-based tagger.