API Documentation ================= .. admonition:: Start Here - Estimation theory and estimator families: :doc:`Longitudinal` - Diagnostics workflow: :doc:`Diagnostics` - Semiparametric estimator narratives: :doc:`Semiparametrics` .. contents:: On This Page :local: :depth: 2 Overview -------- This page is the reference hub for public APIs. Canonical mathematical and workflow discussions remain in the domain pages; this page links to curated reference surfaces for direct function/class lookup. Longitudinal Estimator APIs --------------------------- .. list-table:: :header-rows: 1 :widths: 25 35 20 20 * - API Group - When To Use - Required Inputs - Key Outputs * - RKHS (`RKHSIV*`, `RKHS2IV*`) - Kernel-based sequential/joint NPIV with regularization and CV. - `A, B, C, C', Y` - Fitted bridge/structural predictors * - Neural Network (`AGMM*`, `AGMM2L2`) - Adversarial minimax estimation with flexible representation learning. - `A, B, C, C', Y` tensors/arrays - Learned predictors and training diagnostics * - Ensemble / Random Forest (`Ensemble*`) - Oracle-style ensemble approximations for minimax objectives. - `A, B, C, C', Y` - Ensemble estimators for `g`/`h` * - Sparse/Regularized Linear + TSLS - Interpretable baselines and constrained optimization settings. - Matrix covariates + outcomes - Coefficients and predictions .. toctree:: :maxdepth: 2 mliv_overview Diagnostics APIs ---------------- Canonical diagnostics reference: :doc:`diagnostics/Universal` Semiparametric APIs ------------------- .. list-table:: :header-rows: 1 :widths: 30 35 20 15 * - API - Target - Core Inputs - Main Return * - `DML_npiv` - NPIV functional inference - `Y, D, Z, W` (+ options) - `theta, var, ci` * - `DML_mediated` - Mediation estimands - mediated DGP blocks - effect estimates + CI * - `DML_longterm` - Long-term treatment effects - long-term/surrogacy blocks - effect estimates + CI .. toctree:: :maxdepth: 2 module_overview Related Pages ------------- - :doc:`Longitudinal` - :doc:`Diagnostics` - :doc:`Semiparametrics`