This dissertation from Trinity College – School of Computer Science and Statistics examines the impact of financial media sources on the sentiment of the investing public. The research analyzes investor emotions and views within the United States, the United Kingdom, and India to establish if the impact of media outlets has any measurable influence on international financial markets.
To tackle this study, researchers developed a natural language processing system to extract sentiment from the finance, economics, and trade media. Then using vector auto-regressive methods and a multivariate time-series analysis, they model the impact of investor sentiment on the market return. The study features potential tools that researchers can use to handle a project of this scale, including the Lexalytics Intelligence Platform.
The research argues that sentiment has an episodic and time-varying impact on performance and that their results are statistical significance across each of the three markets. Click the button below to read the full article now: