We present a tool which analyzes annotations on a document to infer collective sentiments of annotators. Annotations may include comments, notes, observation, explanation or question, help etc. Comments are used for evaluative purpose where as others are used either for summarization or for expansion. Further, these comments may be on another annotation, not on the original document and referred as meta-annotations. Collective sentiments of annotators are classified as positive, negative or neutral based on sentiments of words found in annotations. All annotations may not get equal weightage. If an annotation has higher number of meta-annotations on it, it is assigned higher weight. If a comment is on another annotation and negates the sentiments of previous annotator, then the weightage of that annotation is either reduced or annotation is excluded from inference. Our tool computes collective sentiments of annotators in two steps. In first step, it computes sentiment scores of all annotations. In second steps, it computes weighted average of sentiment scores of annotation to obtain the collective sentiments. We demonstrate the use of tool on research papers.
Weitere Kapitel dieses Buchs durch Wischen aufrufen
Bitte loggen Sie sich ein, um Zugang zu diesem Inhalt zu erhalten
Sie möchten Zugang zu diesem Inhalt erhalten? Dann informieren Sie sich jetzt über unsere Produkte:
- Annotation Based Collective Opinion Analysis
B. D. Chaudhary
- Springer Berlin Heidelberg
Neuer Inhalt/© ITandMEDIA