For the last four years, a startup based in the north-eastern French city of Metz, called SESAMm, has been collecting data from thousands of online information sources, such as social network posts, blogs, forums etc., in a variety of languages on different continents. The company staff then analyze this information using natural-language processing (NLP) techniques, which combine Artificial Intelligence and linguistic analysis, plus applied mathematics, to predict fluctuations in stock market prices, especially movements in stock indices. This novel approach to market finance decisions has already caught on with a number of banks and hedge funds (unlisted investment funds designed mainly for speculative investing) and is also targeted at asset managers and insurers. In addition to its website entitled L'Humeur des Marchés ('Market Sentiment'), a platform that enables you to visualize the emotions behind the data for financial asset and portfolio management purposes, the French startup has also designed two other stock prediction products – a 'premium services' offer for customized investment-related projects and a high-frequency trading tool for cryptocurrencies. Two years ago, SESAMm was adjudged among the winners of the Connected Finance Social Media challenge run by BNP Paribas. Now, following two successful fund-raising rounds, the young company, which has establishments in France and Luxembourg, is planning to set up a new subsidiary abroad.
By Marie-Eléonore Noiré