An Artificial Intelligence-Based Industry Peer Grouping System
Jul 15, 2022
In this paper published in the Journal of Financial Data Science, we collaborated with MIT researchers to develop a data-driven industry peer-grouping system that clustered similar companies at different levels of granularity. We also developed an interactive tool to visualize clusters and nearest neighbors of companies. Artificial-intelligence techniques can extract features from relevant data sources and learn relationships that can identify companies that are similar in terms of risk-return profile for the out-of-sample period. Historical returns correlation, Global Industry Classification Standard (GICS®) classification, 10-K reports, and fundamental factors such as size, momentum and debt-to-asset ratio contributed the most in predicting the similarity of companies.1 ©2022 With Intelligence. Republished with permission from: Bonne, George, et al. 2022. “An Artificial Intelligence-Based Industry Peer Grouping System.” Journal of Financial Data Science 4, no. 2 (Spring).
1GICS is a standard jointly developed by MSCI and S&P Global Market Intelligence.