High-Dimensional Inference

high-dimensional-statisticslassovariable-selectionpost-selection-inferencemultiple-testingfalse-discovery-ratesparsity

Definition

High-dimensional inference concerns estimation and — especially — valid uncertainty quantification when the number of parameters pp is comparable to or far exceeds the sample size nn (pnp\gg n). In the sparse linear model Y=Xβ+εY=X\beta+\varepsilon with only a few nonzero coefficients, point estimation is handled by penalized methods such as the Lasso, but producing trustworthy p-values and confidence intervals for individual coefficients requires additional machinery, because selecting and testing on the same data invalidates classical standard errors (Bühlmann-Kalisch-Meier 2014).

Key Ideas

How It Works

  1. Fit a sparse estimator (Lasso, scaled Lasso) to obtain a candidate active set and a noise-level estimate.
  2. Choose an inference strategy: multi-sample-splitting or stability selection for robust selection with error control, or the de-sparsified Lasso for coordinate-wise p-values and confidence intervals.
  3. Apply a multiple-testing correction (FWER or FDR) across the many coefficients tested.
  4. Report selected variables with calibrated significance, acknowledging dependence on the sparsity/design assumptions.

Why It Matters

Open Questions

Related