Semantic Sentiment Analysis of Twitter

Abstract

Abstract. Sentiment analysis over Twitter offer organisations a fast and effective way to monitor the publics' feelings towards their brand, business, directors, etc. A wide range of features and methods for training sentiment classifiers for Twitter datasets have been researched in recent years with varying results. In this paper, we introduce a novel approach of adding semantics as additional features into the training set for sentiment analysis. For each extracted entity (e.g. iPhone) from tweets, we add its semantic concept (e.g. "Apple product") as an additional feature, and measure the correlation of the representative concept with negative/positive sentiment. We apply this approach to predict sentiment for three different Twitter datasets. Our results show an average increase of F harmonic accuracy score for identifying both negative and positive sentiment of around 6.5% and 4.8% over the baselines of unigrams and part-of-speech features respectively. We also compare against an approach based on sentiment-bearing topic analysis, and find that semantic features produce better Recall and F score when classifying negative sentiment, and better Precision with lower Recall and F score in positive sentiment classification.

SentiCircles for Contextual and Conceptual Semantic Sentiment Analysis of Twitter

“…In our previous work we showed that incorporating general conceptual semantics (e.g., "president", "company") into supervised classifiers improved sentiment accuracy [18]. SenticNet [8], 3 is a conceptbased lexicon for sentiment analysis.…”

“…In this section we describe the addition of conceptual semantics into the SentiCircle representation. As in our previous work [18], AlchemyAPI 6 came first amongst the set of entity extractors we tested on Twitter. Here we use AlchemyAPI again to extract all named entities in tweets with their associated concepts.…”

“…Here we use AlchemyAPI again to extract all named entities in tweets with their associated concepts. We add the concepts into the SentiCircle representation using the Semantic Augmentation method [18], where we add the semantic concepts to the original tweet before applying our representation model (e.g., "headache" and its concept "Health Condition" will appear together in the SentiCircle). Also note that each extracted concept will be represented by a SentiCircle in order to compute its overall sentiment.…”

Semantic Patterns for Sentiment Analysis of Twitter

“…However, it has been argued that sentiment in text is not always associated with individual words, but instead, through relations and dependencies between words, which often formulate sentiment [16]. In previous work, these relations are usually complied as a set of syntactic patterns (i.e., Part-of-Speech patterns) [25,16,24], common sense concepts [6], semantic concepts [21,8], or statistical topics [20,11].…”

“…To extract the entities and their associated concepts in our datasets we use AlchemyAPI, 5 which we have previously evaluated its semantic extraction performance on Twitter data [21]. The number of extracted concepts in each dataset is listed in Table 3.…”