Design and Evaluation of SentiEcon: a fine-grained Economic/Financial Sentiment Lexicon from a Corpus of Business News

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 5065–5072 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 5065 Design and Evaluation of SentiEcon: a fine-grained Economic/Financial Sentiment Lexicon from a Corpus of Business News Antonio Moreno-Ortiz, Javier Fern´andez-Cruz, Chantal P´erez-Hern´andez Universidad de M´alaga (Spain) Pontificia Universidad Cat´olica del Ecuador Sede Esmeraldas (Ecuador) {amo,fernandezcruz,mph}@uma.es Abstract In this paper we present, describe, and evaluate SentiEcon, a large, comprehensive, domain-specific computational lexicon designed for sentiment analysis applications, for which we compiled our own corpus of online business news. SentiEcon was created as a plug-in lexicon for the sentiment analysis tool Lingmotif, and thus it follows its data structure requirements and presupposes the availability of a general-language core sentiment lexicon that covers non-specific sentiment-carrying terms and phrases. It contains 6,470 entries, both single and multi-word expressions, each with tags denoting their semantic orientation and intensity. We evaluate SentiEcon’s performance by comparing results in a sentence classification task using exclusively sentiment words as features. This sentence dataset was extracted from business news texts, and included certain key words known to recurrently convey strong semantic orientation, such as “debt”, “inflation” or “markets”. The results show that performance is significantly improved when adding SentiEcon to the general-language sentiment lexicon. Keywords: sentiment analysis, opinion mining, financial texts, lexicons, language resources 1. Introduction The analysis of public opinion on globally relevant enti- ties has consistently been a matter of great import, espe- cially since the advent of market research, starting in the second half of the 20th century. As the number and vari- ety of information resources has grown exponentially due to the consolidation of the Web 2.0 in the last decade, we face the challenge of transforming the richness of this vast amount of information into computationally tractable data. From an economic point of view, there has always been a demand for information on opinions related to financial in- stitutions such as central banks, or the appraisal of the eco- nomic state-of-affairs, such as the Monthly Michigan Index of Consumer Sentiment in the USA. Generally funded by either public or private institutions, these studies were car- ried out employing traditional surveying techniques, with data analysis methods that are not always transparent and, in many cases, without the possibility to access the origi- nal data, which would allow replication of results or find alternative interpretations. A newer trend, however, has been used in recent years which employs data and text mining techniques. Senti- ment Analysis can no doubt be seen as an alternative to traditional surveying methods, and has been considered as a more dynamic and robust solution to the analysis of on- line text, such as news, user-generated content (UGC) or financial reports (Batra and Rao, 2010; Yu et al., 2013). Despite the formidable body of research generated on the topic of sentiment analysis (see Liu (2015) for a concise state-of-the-art), SA strategies are still far from perfect as systems fall short of accounting for the vast complexity …

Content truncated to 500 words