Class Inheritance Diagrams

The diagrams below highlight the essential inheritance relationships in PyTUQ — groups where classes form multi-level hierarchies.

Fitting

fitbase is the root fitting class, extended by gp (Gaussian‑process fitting) and lreg (linear regression).

Inheritance diagram of pytuq.fit.fit, pytuq.fit.gp, pytuq.lreg.lreg

Linear Regression

lreg inherits from fitbase and is specialised into several solvers: lsq (least squares), anl (analytical Bayesian), bcs (Bayesian compressive sensing), opt (optimisation‑based), and lreg_merr (model‑error regression).

Inheritance diagram of pytuq.lreg.lreg, pytuq.lreg.anl, pytuq.lreg.bcs, pytuq.lreg.opt, pytuq.lreg.merr

MCMC Samplers

MCMCBase is the common ancestor of the three samplers: AMCMC (adaptive Metropolis), HMC (Hamiltonian Monte Carlo), and MALA (Metropolis‑Adjusted Langevin Algorithm).

Inheritance diagram of pytuq.minf.calib, pytuq.minf.mcmc

Priors

Prior is specialised into Prior_uniform and Prior_normal.

Likelihoods

Likelihood branches into several concrete likelihood models.

Random Variables

MRV is the multivariate random variable base class, extended by GMM, Mixture, Inverse, and several 1‑D distributions. PCRV (polynomial‑chaos random variable) has its own sub‑hierarchy.

Inheritance diagram of pytuq.rv.mrv, pytuq.rv.pcrv

Functions

Function is the base for all test functions, operators, and benchmark functions in PyTUQ.

Inheritance diagram of pytuq.func.func, pytuq.func.poly, pytuq.func.genz, pytuq.func.oper, pytuq.func.toy

Global Sensitivity Analysis

SensMethod is the root for all GSA methods.

Linear Dimensionality Reduction

LinRed is extended by KLE (Karhunen–Loève) and SVD.

Inheritance diagram of pytuq.linred.linred, pytuq.linred.kle, pytuq.linred.svd

Optimisers

OptBase is the optimiser base class, with gradient‑descent variants (GD, SGD, Adam), particle‑swarm (PSO), and a SciPy wrapper.

Inheritance diagram of pytuq.optim.optim, pytuq.optim.gd, pytuq.optim.pso, pytuq.optim.sciwrap