Characterizing debate performance via aggregated twitter sentiment

Abstract. Lexicon-based approaches to Twitter sentiment analysis are gaining much popularity due to their simplicity, domain independence, and relatively good performance. These approaches rely on sentiment lexicons, where a collection of words are marked with fixed sentiment polarities. However, words' sentiment orientation (positive, neural, negative) and/or sentiment strengths could change depending on context and targeted entities. In this paper we present SentiCircle; a novel lexicon-based approach that takes into account the contextual and conceptual semantics of words when calculating their sentiment orientation and strength in Twitter. We evaluate our approach on three Twitter datasets using three different sentiment lexicons. Results show that our approach significantly outperforms two lexicon baselines. Results are competitive but inconclusive when comparing to state-of-art SentiStrength, and vary from one dataset to another. SentiCircle outperforms SentiStrength in accuracy on average, but falls marginally behind in F-measure.

Section: Datasetmentioning

confidence: 99%

SentiCircles for Contextual and Conceptual Semantic Sentiment Analysis of Twitter

Saif

Fernández

et al. 2014

“…To validate our above observation, we analyse the human sentiment votes on the 58 entities in STS-Gold dataset. 7 Figure 5 shows entities under patterns 2,5 and 6 along with number of times they receive negative, positive and neutral sentiment in tweets according to the three human coders. We observe that entities in patterns 2 and 6 occur very infrequently in tweets, yet with consistent sentiment.…”

Section: Within-pattern Sentiment Consistencymentioning

confidence: 99%

Semantic Patterns for Sentiment Analysis of Twitter

Saif

Fernández

et al. 2014

Abstract. Most existing approaches to Twitter sentiment analysis assume that sentiment is explicitly expressed through affective words. Nevertheless, sentiment is often implicitly expressed via latent semantic relations, patterns and dependencies among words in tweets. In this paper, we propose a novel approach that automatically captures patterns of words of similar contextual semantics and sentiment in tweets. Unlike previous work on sentiment pattern extraction, our proposed approach does not rely on external and fixed sets of syntactical templates/patterns, nor requires deep analyses of the syntactic structure of sentences in tweets. We evaluate our approach with tweet-and entity-level sentiment analysis tasks by using the extracted semantic patterns as classification features in both tasks. We use 9 Twitter datasets in our evaluation and compare the performance of our patterns against 6 state-of-the-art baselines. Results show that our patterns consistently outperform all other baselines on all datasets by 2.19% at the tweet-level and 7.5% at the entity-level in average F-measure.

“…Shamma and Diakopoulos [15] showed that the social structure and the conversational content of tweets can provide insight into a media event's structure and semantic content: quantitatively through activity peaks and qualitatively through keyword mining. Diakoloulos et al [16], after having collected tweets during the U.S. State of the Union presidential address in 2010, used them to annotate a video of the event.…”

Section: Crowdsourcing Media Annotation and Motivationmentioning

confidence: 99%

“…Instead, for PT [34], we needed to find a solution that would work in multiple and mixed languages, was simple to deploy, and would work on a variety of specialized corpora like philosophy, aesthetics, or design. Diakopoulos and Shamma [15] used another approach for sentiment classification: they used Amazon Mechanical Turk to perform hand-annotated sentiment classification on tweets. Turkers were compensated $0.05 per ten tweets analyzed.…”

Section: Tweet Content Analysismentioning

confidence: 99%

See 1 more Smart Citation

PolemicTweet: Video Annotation and Analysis through Tagged Tweets

Huron¹,

Isenberg²,

Fekete³

2013

Abstract. We present PolemicTweet a system with an encompassing, economic, and engaging approach to video tagging and analysis. Annotating and tagging videos manually is a boring and time-consuming process. Yet, in the last couple of years the audiences of events-such as academic conferences-have begun to produce unexploited metadata in the form of micropost activities. With PolemicTweet we explore the use of tagged microposts for both video annotation and browsing aid. PolemicTweet is a system 1) to crowd source conference video tagging with structured sentiment metadata, 2) to engage audiences in a tagging process, and 3) to visualize these annotations for browsing and analyzing a video. We describe the the system and its components as well as the results from a one-year live deployment in 27 different events.