M-Eco D3.2 - Semantic Annotator

named entity recognition, semantic annotation, Twitter

This deliverable deals with semantic annotation of texts collected within the M-Eco project. We describe the annotation process which involves various named entity recognition. A special attention is paid to the key elements for generating signals - geographical names, temporal expressions, diseases and symptoms. The designed and implemented annotation tool is also evaluated on several datasets developed within the project. It provides state-of-the-art performance in terms of annotation accuracy as well as excellent scalability - it is able to on-line process the continuous stream of texts coming through the M-Eco system.

This deliverable deals with semantic annotation of texts collected within the M-Eco project. We describe the annotation process which involves various named entity recognition. A special attention is paid to the key elements for generating signals - geographical names, temporal expressions, diseases and symptoms. The designed and implemented annotation tool is also evaluated on several datasets developed within the project. It provides state-of-the-art performance in terms of annotation accuracy as well as excellent scalability - it is able to on-line process the continuous stream of texts coming through the M-Eco system.