Automated Detection of Financial Events in News Text Defended on Thursday, 11 December 2014
Today’s financial markets are inextricably linked with financial events like acquisitions, profit announcements, or product launches. Information extracted from news messages that report on such events could hence be beneficial for financial decision making. The ubiquity of news, however, makes manual analysis impossible, and due to the unstructured nature of text, the (semi-)automatic extraction and application of financial events remains a non-trivial task. Therefore, the studies composing this dissertation investigate 1) how to accurately identify financial events in news text, and 2) how to effectively use such extracted events in financial applications.
Based on a detailed evaluation of current event extraction systems, this thesis presents a competitive, knowledge-driven, semi-automatic system for financial event extraction from text. A novel pattern language, which makes clever use of the system’s underlying knowledge base, allows for the definition of simple, yet expressive event extraction rules that can be applied to natural language texts. The system’s knowledge-driven internals remain synchronized with the latest market developments through the accompanying event-triggered update language for knowledge bases, enabling the definition of update rules.
Additional research covered by this dissertation investigates the practical applicability of extracted events. In automated stock trading experiments, the best performing trading rules do not only make use of traditional numerical signals, but also employ news-based event signals. Moreover, when cleaning stock data from disruptions caused by financial events, financial risk analyses yield more accurate results. These results suggest that events detected in news can be used advantageously as supplementary parameters in financial applications.
Financial events, news processing, event extraction, information extraction, natural language processing, semantic web, ontologies, event-driven systems, algorithmic trading, value at risk