This repository contains the files composing the life! gold standard corpus and further improvements. It has been created by PlataformaLifeUA.

Files are stored as XML following Gate standoff format. The root element of each file is <GateDocument>. Every document contains:

• <TextWithNodes> to store the text of the annotated document. <Node> indicates that this position is either the initial and final character offset of an annotation.
• For files which belong to gold standard corpus, annotations are organised in several <AnnotationSet>. Each set includes detailed anotations carried out by different authors. Specifically:
  • 15 markups and 80 markups contains manual annotations performed by human raters (with identifiers 15 and 80).
  • consensus markups gathers the gold standard annotations that were included by consensus of the Life Platform team.
  • LingPipe markups includes annotations included by the tokenizer and sentence splitter from LingPipe pluging.

At this moment, Life! corpus contains 8043 tokens and 536 sentences in 102 document. Each document is a message extracted from Surface or Deep Web Social Networks. Each message contains three properties manually annotated, namely: Language, Alert level and Type of message.

Next tables provide more details of each property:

• Gate Developer 8.2: The annotation tool used

This resource is the output of the following research projects: (i) Análisis de Sentimientos Aplicado a la Prevención del Suicidio en las Redes Sociales (ASAP) funded by Fundación BBVA, (ii) Representación canónica y transformaciones de los textos aplicado a las Tecnologías del Lenguaje Humano (RESCATA - Ref. TIN2015-65100-R) and (iii) Reconocimiento de Entidades Digitales: Enriquecimiento y Seguimiento (REDES - Ref. TIN2015-65136-C2-2-R)