New method detects hidden patterns in big data for better predictions.
The article presents three essays on large panel data models with cross-sectional dependence. The researchers developed a test to detect cross-sectional dependence and estimate error terms efficiently. They also analyzed nonparametric dynamic panel data models with interactive fixed effects, proposing sieve estimation for nonparametric functions and a consistent bias-corrected estimator. The key findings include establishing the asymptotic normality of the test statistic and deriving the convergence rate for the sieve estimator.