funSentiment at SemEval-2017 Task 5: Fine-Grained Sentiment Analysis on Financial Microblogs Using Word Vectors Built from StockTwits and Twitter

Proceedings of the 11th International Workshop on Semantic Evaluations (SemEval-2017), pages 852–856, Vancouver, Canada, August 3 - 4, 2017. c⃝2017 Association for Computational Linguistics funSentiment at SemEval-2017 Task 5: Fine-Grained Sentiment Anal- ysis on Financial Microblogs Using Word Vectors Built from StockTwits and Twitter Quanzhi Li, Sameena Shah, Armineh Nourbakhsh, Rui Fang, Xiaomo Liu Research and Development Thomson Reuters 3 Times Square, NYC, NY 10036 {quanzhi.li, sameena.shah, armineh.nourbakhsh, rui.fang, xiaomo.liu}@thomsonreuters.com Abstract This paper describes the approach we used for SemEval-2017 Task 5: Fine-Grained Sentiment Anal- ysis on Financial Microblogs. We use three types of word embeddings in our algorithm: word embeddings learned from 200 million tweets, sentiment-specific word embeddings learned from 10 million tweets us- ing distance supervision, and word embeddings learned from 20 million StockTwits messages. In our approach, we also take the left and right context of the target company into consideration when generating polarity prediction features. All the features generated from different word embeddings and contexts are in- tegrated together to train our algorithm. 1 Introduction Domain specific Sentiment Analysis has re- ceived much attention recently. The financial domain is a high-impact use case for Sentiment Analysis because it has been shown that senti- ments and opinions can affect market dynamics [9, 48]. Given the link between sentiment and market dynamics, the analysis of public senti- ment becomes a powerful method to predict the market reaction. One main source of public sen- timent is social media, such as Twitter and StockTiwts. In this paper, we describe our approach for SemEval-2017 Task 5: Fine-Grained Sentiment Analysis on Financial Microblogs (Cortis et al., 2017). The task is: given a microblog message, predict the sentiment score for each of the com- panies/stocks mentioned. Sentiment values need- ed to be floating point values within the range of -1 (very negative/bearish) to 1 (very positive/ bullish), with 0 designating neutral sentiment. Our approach uses word embeddings (WE- Twitter) learned from general tweets, sentiment specific word embeddings (SSWE) learned from distance supervised tweets, and word embeddings learned from StockTwits messages (WE-StockTwits). Message or sentence level sentiment classifi- cation has been studied by many previous works (Go et al., 2009; Mohammand et al., 2013; Pang et al., 2002; Liu, 2012; Tang et al., 2014), but there are few studies on target-dependent, or en- tity level, sentiment prediction (Jiang et al., 2011; Dong et al., 2014; Vo and Zhang, 2015). A target entity in a message does not necessarily have the same polarity type as the message, and different entities in the same message may have different polarities. For example, in the tweet “iPhone is better than Blackberry”, the two named entities, iPhone and Blackberry, will have different sentiment polarities. Recent studies have focused on learning features directly from tweet text. One approach is to generate sentence representations from word embeddings. Several word embedding generation algorithms have been proposed in previous studies (Collobert et al., 2011; Mikolov et al., 2013). Using the gen- eral word embeddings directly in sentiment anal- ysis is not effective, since …

Content truncated to 500 words