Abstract
Sentiment analysis is a content-analytic investigative framework for researchers, traders and the general public involved in financial markets. This analysis is based on carefully sourced and elaborately constructed proxies for market sentiment and has emerged as a basis for analysing movements in stock prices and the associated traded volume. This approach is particularly helpful just before and after the onset of market volatility. We use an autoregressive framework for predicting the overall changes in stock prices by using investor sentiment together with lagged variables of prices and trading volumes. The case study we use is a small market index (Danish Stock Exchange Index, OMXC 20, together with prevailing sentiment in Denmark, to evaluate the impact of sentiment on OMXC 20. Furthermore, we introduce a rather novel and quantitative sentiment proxy, that is the use of the index of a larger market (US S&P 500), to see how the smaller market reacts to changes in the larger market. The use of larger market index is justified on economic/financial grounds in that globalisation has introduced a degree of interdependence, and allow us to explore global influences as a proxy for sentiment. We look at the robustness of our prediction. (Local) Negative sentiment (as articulated in Danish newspapers over a 7 year period (2007–2013), does have an impact on the local markets, but the global market (S&P 500) has an even greater impact.
Original language | English |
---|---|
Title of host publication | Intelligent Data Engineering and Automated Learning – IDEAL 2015 : 16th International Conference Wroclaw, Poland, October 14–16, 2015, Proceedings |
Editors | Konrad Jackowski, Robert Burduk, Krzysztof Walkowiak, Michał Woźniak, Hujun Yin |
Number of pages | 9 |
Volume | 9375 |
Place of Publication | Cham |
Publisher | Springer Science+Business Media |
Publication date | 2015 |
Pages | 466-474 |
ISBN (Print) | 9783319248332 |
ISBN (Electronic) | 9783319248349 |
DOIs | |
Publication status | Published - 2015 |
Event | The 16th International Conference on Intellligent Data Engineering and Automated Learning - Wroclaw University of Technology, Wroclaw, Poland Duration: 14 Oct 2015 → 16 Oct 2015 Conference number: 16 http://ideal2015.pwr.edu.pl/ |
Conference
Conference | The 16th International Conference on Intellligent Data Engineering and Automated Learning |
---|---|
Number | 16 |
Location | Wroclaw University of Technology |
Country | Poland |
City | Wroclaw |
Period | 14/10/2015 → 16/10/2015 |
Internet address |
Series | Lecture Notes in Computer Science |
---|---|
Volume | 9375 |
ISSN | 0302-9743 |
Keywords
- Sentiment analysis
- Equity index markets
- Time series analysis
- Vector Autoregressive
- Granger causality