ECNU at SemEval-2016 Task 4: An Empirical Investigation of Traditional NLP Features and Word Embedding Features for Sentence-level and Topic-level Sentiment Analysis in Twitter

Proceedings of SemEval-2016, pages 256–261, San Diego, California, June 16-17, 2016. c⃝2016 Association for Computational Linguistics ECNU at SemEval-2016 Task 4: An Empirical Investigation of Traditional NLP Features and Word Embedding Features for Sentence-level and Topic-level Sentiment Analysis in Twitter Yunxiao Zhou1, Zhihua Zhang1, Man Lan1,2∗ 1Department of Computer Science and Technology, East China Normal University, Shanghai, P.R.China 2Shanghai Key Laboratory of Multidimensional Information Processing {10122130215, 51131201039}@ecnu.cn, mlan@cs.ecnu.edu.cn∗ Abstract This paper reports our submissions to Task 4, i.e., Sentiment Analysis in Twitter (SAT), in SemEval 2016, which consists of five subtasks grouped into two levels: (1) sentence level, i.e., message polarity classification (subtask A), and (2) topic level, i.e., tweet classification and quantification according to two-point s- cale (subtask B and D) or five-point scale (sub- task C and E). We participated in all these five subtasks. To address these subtasks, we inves- tigated several traditional Natural Language Processing (NLP) features including senti- ment lexicon, linguistic and domain specific features, and word embedding features togeth- er with supervised machine learning methods. Officially released results showed that our sys- tems rank above average. 1 Introduction In recent years, with the emergence of social media, more and more users have shared and obtained in- formation through microblogging websites, such as Twitter. As a result, a huge amount of available data attracts a lot of researchers. SemEval 2016 provides such a universal platform for researchers to explore in the task of Sentiment Analysis in Twitter (Nakov et al., 2016) (Task 4), which includes five subtasks grouped into two levels, i.e., sentence level and top- ic level. Subtask A is a sentence level task aiming at sentiment polarity classification of the whole tweet. The other four subtasks are at topic level, i.e., given one topic, the sentiment polarity of tweets are clas- sified or assigned by a two-point scale (i.e., subtask B and D) and by a five-point scale (i.e., subtask C and E). Specifically, subtask B is to identify the sen- timent polarity label (i.e, Positive and Negative) of tweets with respect to the given topic while subtask D aims at estimating the sentiment distribution of tweets with respect to the given topic. Both subtask B and D are on a two-point scale. The purposes of subtask C and E are similar with that of subtask B and D, except for using a five-point scale, that is, the class labels are of 5 values, i.e, 2, 1, 0, -1 and -2 rep- resenting Very Positive, Positive, Neutral, Negative and Very Negative. Given the character limitations on tweets, senti- ment orientation classification on tweets can be re- garded as a sentence-level sentiment analysis. Many researchers focus on feature engineering to improve the performance of SAT. For example, (Turian et al., 2010; Liu, 2012; Zhang et al., 2006) showed that one-hot representation on n-gram features is a rela- tively strong baseline. Furthermore, (Mohammad et al., 2013) proposed a state-of-the-art model which implemented several sentiment lexicons and a vari- ety of manual features. Apart from the …

Content truncated to 500 words