Data-driven sparse identification of governing PDEs via knockoff filters and multi-criteria trade-offs
Researchers have developed a new data-driven framework called KO-PDE-IDENT for discovering partial differential equations (PDEs) from noisy data. This method uses knockoff filters to control the false discovery rate, addressing issues of multicollinearity that plague traditional sparse regression techniques. The framework integrates SHAP values with recursive feature elimination and a multi-criteria decision-making process to balance accuracy, complexity, and coefficient uncertainty, demonstrating accurate PDE structure recovery in simulations. AI
IMPACT Introduces a novel method for scientific discovery, potentially accelerating research in fields reliant on partial differential equations.