GPstuff Manual

GPstuffManual

User Manual:

Open the PDF directly: View PDF .
Page Count: 77

Scroll down to view the document on your mobile browser.

Introduction
Gaussian process models
- Gaussian process prior
- Conditioning on the observations
Conditional posterior and predictive distributions
- Gaussian observation model: the analytically tractable case
- Laplace approximation
- Expectation propagation algorithm
- Markov chain Monte Carlo
Marginal likelihood given parameters
Marginalization over parameters
- Maximum a posterior estimate of parameters
- Grid integration
- Monte Carlo integration
- Central composite design
Getting started with GPstuff: regression and classification
- Gaussian process regression
- Gaussian process classification
Other single latent models
- Robust regression
  - Student-t observation model
  - Laplace observation model
- Count data
- Log-Gaussian Cox process
- Accelerated failure time survival models
- Derivative observations in GP regression
- Quantile Regression
Multilatent models
- Multiclass classification
- Multinomial
- Cox proportional hazard model
- Zero-inflated negative binomial
- Density estimation and regression
- Input dependent models
  - Input-dependent Noise
  - Input-dependent overdispersed Weibull
- Monotonicity constraint
Mean functions
- Explicit basis functions
Sparse Gaussian processes
- Compactly supported covariance functions
  - Compactly supported covariance functions in GPstuff
- FIC and PIC sparse approximations
  - FIC sparse approximation in GPstuff
  - PIC sparse approximation in GPstuff
- Deterministic training conditional, subset of regressors and variational sparse approximation
  - Variational, DTC and SOR sparse approximation in GPstuff
- Sparse GP models with non-Gaussian likelihoods
- Predictive active set selection for classification
Modifying the covariance functions
- Additive models
  - Additive models in GPstuff
- Additive covariance functions with selected variables
- Product of covariance functions
State space inference
Model assessment and comparison
- Marginal likelihood
- Cross-validation
- DIC
- WAIC
- Model assessment demos
Bayesian optimization
Adding new features in the toolbox
Discussion
Comparison of features in GPstuff, GPML and FBM
Covariance functions
Observation models
Priors
Transformation of hyperparameters
Developer appendix
- lik functions
  - lik_negbin
  - lik_t
- gpcf functions
  - gpcf_sexp
- prior functions
- Optimisation functions
- Other inference methods

Navigation menu