Understanding Markets by Understanding Investors
Investors group stocks by familiar categories—sector, market cap, growth vs. value. Our AI model trained on institutional portfolio data discovers its own groupings—hidden patterns in investor behavior that often explain valuations and price moves better than traditional metrics.
THE SCIENCE BEHIND
Building on Asset Embeddings by Gabaix, Koijen, Richmond & Yogo, The Market Genome Project applies investor embedding methods to public filings data, letting you interactively explore similarity structures, portfolio dynamics, and the interpretable narratives to interpret investors' actions.
Read the ResearchKEY INSIGHTS
Transformer models trained on real-time portfolio data cluster stocks by how institutions actually hold them and investors by the stocks they hold—uncovering themes that traditional categories are not designed to capture.
Every insight is grounded in actual institutional holdings from SEC filings, providing an authentic view of how professional investors structure portfolios in the current market environment.
Earnings call transcripts linked to holdings and rebalancing surface the management narratives and catalysts that drive position changes—connecting the story to the actions.
The Market Genome Project is intended for non-commercial research, educational, and informational use only. It is not investment advice or a recommendation and does not create an advisory or client relationship. AI-generated analysis may be incomplete or inaccurate. Read the Research Notice.
Questions about The Market Genome Project? Contact@MktGen.ai