A Classification-based Approach to Economic Event Detection in Dutch News Text

A Classification-based Approach to Economic Event Detection in Dutch News Text Els Lefever and V´eronique Hoste LT3 Language and Translation Technology Team Ghent University Groot-Brittanni¨elaan 45, 9000 Ghent, Belgium els.lefever, veronique.hoste@ugent.be Abstract Breaking news on economic events such as stock splits or mergers and acquisitions has been shown to have a substantial impact on the financial markets. As it is important to be able to automatically identify events in news items accurately and in a timely manner, we present in this paper proof-of-concept experiments for a supervised machine learning approach to economic event detection in newswire text. For this purpose, we created a corpus of Dutch financial news articles in which 10 types of company-specific economic events were annotated. We trained classifiers using various lexical, syntactic and semantic features. We obtain good results based on a basic set of shallow features, thus showing that this method is a viable approach for economic event detection in news text. 1. Introduction In the financial domain, the way companies are perceived by investors is highly influenced by the news (Engle and Ng, 1993; Tetlock, 2007; Mian and Sankaraguruswamy, 2012). This news can be either linked to macroeconomic factors (like the overall economy and industry conditions), news from the geo-political front or company-specific fac- tors (like the financial condition or announcements about dividend payments or stock-splits). Good news would tend to lift the market, while on the other hand, bad news would tend to dampen the markets growth, although it has also been observed (Engle and Ng, 1993) that there exists an asymmetric impact on the stock market from news, whereby negative news shows a greater impact on volatility than positive news. News might also have different effects on the market return depending on the overall state of the market. Good news in a bullish (positive) market showing confidence may be reacted on very differently from the same piece of news arrived during a bearish (pessimistic) market. Event studies examine the behavior of companies’ stock prices around certain economic events such as dividend announcements, stock splits, mergers and acquisitions, etc. (see MacKinlay (1997) for an overview of event study methods). The vast literature on event studies written over the past several decades has become an important part of financial economics (Kothari and Warner, 2007). In a corporate context, the usefulness of event studies arises from the fact that the magnitude of abnormal performance at the time of an event provides a measure of the (unan- ticipated) impact of this type of event on the wealth of the firms claimholders. Event studies also serve an important purpose in capital market research as a way of testing market efficiency. Studying the impact of specific events on the stock markets, however, is a labor-intensive process. This has prompted the use of text mining techniques for the automatic detection of economic events in news text. Identifying news published about certain events in an auto- matic way enables researchers in the field of event studies to process more data …

Content truncated to 500 words