We are pleased to announce a renewal of our collaboration between Causality Link and the Toulouse School of Economics (TSE). Building upon our initial partnership in June 2018, we are reaffirming our commitment to merging cutting-edge artificial intelligence technology with academic expertise.
The initial collaboration has produced a series of research papers which delve into various aspects of investment research and offer valuable insights that have the potential to reshape the way investment professionals approach their craft.
Read on to learn more about a few of the studies.
Understanding World Economy Dynamics Based on Indicators and Events
By Yasser Abbas and Abdelaati Daouia
This article investigates the portrayal of the world economy in Western media using network data and graph theory techniques. By representing key economic indicators as nodes linked by edges weighted by the frequency of mentions in news articles between January 2018 and January 2022, the study constructs dynamic graphs and uses modularity-based clustering, in the form of the Leiden algorithm, to capture the evolving narrative, particularly around the onset of the Covid-19 pandemic.
Read the whitepaper here.
Parsimonious Wasserstein Text-Mining
By Sebastien Gadat and Stephane Villeneuve
This article presents a new way to analyze text data. It combines two methods, called NMF factorization and supervised clustering with Wasserstein barycenters, to reduce the dimension of the model and make the analysis more efficient. By doing this, the researchers can represent a piece of text as a probability distribution. This means they can understand the meaning of the text while also accounting for any uncertainty, a method that can help to use text information to predict financial data.
Read the whitepaper here.
Stock Market Reaction to News: Do Tense and Horizon Matter?
By Marie Briere, Karen Huynh, Olav Laudy and Sebastien Pouget
Co-authored by Marie Briere, Head of Investor Intelligence and Academic Partnership at Amundi Institute, this paper examines how financial markets react to news about events occurring at different timescales. Examining over 1.7 million news-based signals on over 4,400 US stocks, the findings suggest that markets efficiently incorporate news into prices shortly after its release. Notably, there is a stronger market reaction to news about future events compared to those about the present or past, while news concerning near-future events has a greater impact on prices than news about distant-future events. Additionally, the research found that firms’ size matters: Russell 1000 component firms experienced smaller reactions to news than firms not in the Russell 1000.
For this study, researchers from Amundi Institute and TSE used the Causality Link platform to extract the informational content of firm-specific news. Through both machine learning and symbolic techniques, the platform retrieved daily news and worked to understand different elementary concepts, particularly identified KPIs (for example: earnings, sales, etc.), trends and the tense of news.
Read the whitepaper here.
Learn More
We invite you to explore our website to learn more about Causality Link and to contact us to arrange a conversation. If you would like to sign up for a free trial of our alerts, delivered via email, you may do so at this link.
For more information on Causality Link’s partnership with TSE, you can visit this link, and visit this link for an interview with CEO Pierre Haren discussing the partnership.
