Market surveillance systems (MSSs) are increasingly used to monitor trading activities in financial markets to maintain market integrity. Existing MSSs primarily focus on statistical analysis of market activity data and largely ignore textual market information, including, but not limited to, news reports and various social media. As suggested by both theoretical explorations in finance and prevailing market surveillance practice, unstructured market information holds major yet underexplored opportunities for surveillance. In this paper, we propose a news analysis approach with the help of commonsense knowledge to assess the risk of suspicious transactions identified in market activity analysis. Our approach explicitly models semantic relations between transactions and news articles and provides semantic references to words in news articles. We conducted experiments using data collected from a real-world market and found that our proposed approach significantly outperforms the existing methods, which are based on transaction characteristics or traditional textual analysis methods. Experiments also show that the performance advantage of the proposed approach mainly comes from the modeling of news-transaction relationships. The research contributes to the market surveillance literature and has significant practical implications.
Keywords: market surveillance; text mining; commonsense knowledge; business intelligence; intelligent financial systems
History: Accepted by Ram Ramesh, Area Editor for Knowledge Management and Machine Learning; received November 2013; revised September 2014, May 2015; accepted September 2015. Published online March 14, 2016.
Financial markets, a cornerstone of modern economy, require that participants trade in a fair manner (Comerton-Forde and Rydge 2006). However, there have always been attempts to impair market integrity for economic benefit, such as illegal insider trading and market manipulation. To combat such market abuses, market surveillance systems (MSSs) have been developed to monitor market activities and identify suspicious transactions for investigation (Lucas 1993). MSSs play a central role in preventing misconduct and violation of trading regulations, which goes beyond the direct need of regulatory bodies. Since the 2008 economic crisis and several trading misconduct cases, in which some traders caused their companies huge financial losses, financial firms have been eager to incorporate MSSs to monitor internal activities for risk mitigation (Rodier 2011, Steinert-Threlkeld 2011).
In the literature, studies on MSSs remain sparse. Part of the reason is the sensitive and proprietary nature of the surveillance algorithms (Lucas 1993). To the best of our knowledge, the prevailing approach in MSSs focuses on analyzing changes of market activities, such as prices, trading volumes, or transaction profits (Lucas 1993, Mangkorntong and Rabhi 2009). Whereas such analysis provides useful prescreening functions, it does not fully meet surveillance specialists' needs. In practice, surveillance specialists increasingly rely on market information, such as public news and announcements, mostly in textual format, to judge the risk of suspicious transactions identified through market activity analysis.
From a theoretical perspective, there is a long chain of studies showing that news and other market information are relevant to market activities. As argued in market efficiency studies, properly working financial markets will react to public information in an efficient manner (Green 2004). Market activities should be temporally consistent with...