Sentiment Lexicon Adaptation with Context and Semantics for the Social Web

Tracking #: 1437-2649

Hassan Saif
Miriam Fernandez
Leon Kastler
Harith Alani

Responsible editor: 
Guest Editors Social Semantics 2016

Submission type: 
Full Paper
Sentiment analysis over social streams offers governments and organisations a fast and effective way to monitor the publics' feelings towards policies, brands, business, etc. General purpose sentiment lexicons have been used to compute sentiment from social streams, since they are simple and effective. They calculate the overall sentiment of texts by using a general collection of words, with predetermined sentiment orientation and strength. However, words' sentiment often vary with the contexts in which they appear, and new words might be encountered that are not covered by the lexicon, particularly in social media environments where content emerges and changes rapidly and constantly. In this paper, we propose a lexicon adaptation approach that uses contextual as well as semantic information extracted from DBPedia to update the words' weighted sentiment orientations and to add new words to the lexicon. We evaluate our approach on three different Twitter datasets, and show that enriching the lexicon with contextual and semantic information improves sentiment computation by 3.4% in average accuracy, and by 2.8% in average F1 measure.
Full PDF Version: 


Solicited Reviews:
Click to Expand/Collapse
Review #1
Anonymous submitted on 02/Sep/2016
Review Comment:

The authors took into account my comments about significance testing and explained in detail the tests, the null hypotheses, and the correction of p-values.

The authors also improved the references to support the claim that SentiStrength is the state of the art. The authors also clarified the SentiCircle model and the relation between the rules they apply and SentiStrength's. Similarly, they improved the presentation of results as I suggested.

The results and model behind the paper are still not impressive, but at least the methodology is sound and the evidence is good enough to consider this some incremental contribution to the field.

Review #2
Anonymous submitted on 05/Sep/2016
Review Comment:

This work tackles a problem inherent to sentiment lexicons, i.e. the staticness of sentiment values, in a novel way, by using co-occurrence probabilities and leveraging DBPedia. The results are highly promising and the work is presented neatly and easy to understand. Thus, the three dimensions originality, significance of results and quality of writing are fulfilled. The authors also addressed all of my concerns, thus I suggest to accept the paper for publication.