Meta-level sentiment models for big social data analysis

From Brede Wiki
Jump to: navigation, search
Paper (help)
Meta-level sentiment models for big social data analysis
Authors: Felipe Bravo-Marquez, Marcelo Mendoza, Barbara Poblete
Citation: Knowledge-Based Systems missing volume : missing pages. 2014
Database(s):
DOI: 10.1016/j.knosys.2014.05.016.
Link(s): http://www.cs.waikato.ac.nz/~fjb11/publications/KBS2014.pdf
Search
Web: Bing Google Yahoo!Google PDF
Article: BASE Google Scholar PubMed
Restricted: DTU Digital Library
Other: NIF
Services
Format: BibTeX
Extract: Talairach coordinates from linked PDF: CSV-formated wiki-formated

Meta-level sentiment models for big social data analysis describes an evaluation of sentiment analysis word list and algorithms. Among the word list are AFINN.

Features extracted with the word lists are used to train classifiers. These classifiers perform well.

The sentiment analysis methods are compared on several sentiment-labeled data sets SemEval, Sanders and STS.

[edit] Method

The approached used are: SentiStrength, Sentiment140, OpinionFinder, Liu Lexicon, NRC-emotion, AFINN, SentiWordNet, SenticNet.

[edit] Related papers

  1. Combining strengths, emotions and polarities for boosting Twitter sentiment analysis
Personal tools