Sentiment Polarity Identification in Financial News: A Cohesion-based Approach

Proceedings of the 45th Annual Meeting of the Association of Computational Linguistics, pages 984–991, Prague, Czech Republic, June 2007. c⃝2007 Association for Computational Linguistics Sentiment Polarity Identification in Financial News: A Cohesion-based Approach Ann Devitt School of Computer Science & Statistics, Trinity College Dublin, Ireland Ann.Devitt@cs.tcd.ie Khurshid Ahmad School of Computer Science & Statistics, Trinity College Dublin, Ireland Khurshid.Ahmad@cs.tcd.ie Abstract Text is not unadulterated fact. A text can make you laugh or cry but can it also make you short sell your stocks in company A and buy up options in company B? Research in the domain of finance strongly suggests that it can. Studies have shown that both the informational and affective aspects of news text affect the markets in profound ways, im- pacting on volumes of trades, stock prices, volatility and even future firm earnings. This paper aims to explore a computable metric of positive or negative polarity in financial news text which is consistent with human judgments and can be used in a quantita- tive analysis of news sentiment impact on fi- nancial markets. Results from a preliminary evaluation are presented and discussed. 1 Introduction Research in sentiment analysis has emerged to ad- dress the research questions: what is affect in text? what features of text serve to convey it? how can these features be detected and measured automati- cally. Sentence and phrase level sentiment analy- sis involves a systematic examination of texts, such as blogs, reviews and news reports, for positive, negative or neutral emotions (Wilson et al., 2005; Grefenstette et al., 2004). The term “sentiment analysis” is used rather differently in financial eco- nomics where it refers to the derivation of market confidence indicators from proxies such as stock prices and trading volumes. There is a tradition going back to the Nobel Sveriges–Riksbank Laure- ates Herbert Simon (1978 Prize) and Daniel Kah- neman (2002 Prize), that shows that investors and traders in such markets can behave irrationally and that this bounded rationality is inspired by what the traders and investors hear from others about the con- ditions that may or may not prevail in the markets. Robert Engle (2003 Prize) has given a mathematical description of the asymmetric and affective impact of news on prices: positive news is typically related to large changes in prices but only for a short time; conversely the effect of negative news on prices and volumes of trading is longer lasting. The emergent domain of sociology of finance examines financial markets as social constructs and how communica- tions, such as e-mails and news reports, may be loaded with sentiment which could distort market trading (MacKenzie, 2003). It would appear that news affects the markets in profound ways, impacting on volumes of trade, stock returns, volatility of prices and even future firm earnings. In the domain of news impact analy- sis in finance, in recent years the focus has expanded from informational to affective content of text in an effort to explain the relationship between text and the markets. All text, …

Content truncated to 500 words