Project First, the European Union-backed research effort to find a way of extracting and analysing financial services-related sentiment from social media networks, has launched its first running prototype.
Launched in late 2010, the three year project aims to harness artificial intelligence to find information on the Internet that can then be used to support financial decision makers.
It is being carried out by a consortium including Banca Monte dei Paschi di Siena, Boerse Stuttgart, Interactive Data b-next and the University of Hohenheim, who have a EUR4.6 million budget co-funded by the EU.
Unveiling their first working prototype at a conference in Milan, the consortium revealed that, applied to Twitter or textual data extracted from blogs, it is currently capable of extracting sentiment and relating it to stock price movements.
"Results indicate a strong positive relationship between sentiment and trading volume," say the group.
In addition, by monitoring social chatter among market participants, the technology can detect scenarios of financial market abuse, such as attempts to manipulate prices via false information release, a feature which could help improve supervisory and regulatory oversight. In the future, the system will extend to reputational risks.
Wolfgang Fabisch, CEO, b-next, says: "Recent news has made it clear that colleagues communicate their knowledge and opinion via email, chat rooms and forums. As channels of communication continue to expand it is very possible that sensitive information could also be communicated via social mediums."
A final prototype of the system is expected to be released by the end of the project in September 2013.
Introduction to the First project: