Package index • magp

Fit and predict

Fit either mapping model and predict outcomes with uncertainty.

magp2d_fit(): Fit a MaGP model with a two-dimensional sequence map
magpfull_fit(): Fit a MaGP model with a full sequence map
predict(<magp2d>) predict(<magpfull>): Predict outcomes from a fitted MaGP model
magp2d_rmse(): Calculate root-mean-squared prediction error

Build and assess quantitative-sequence starting designs.

magp_initial_design(): Construct a quantitative-sequence initial design
magp_sequence_design(): Construct the sequence portion of an initial design
magp_quantitative_design(): Construct the quantitative portion of an initial design
magp_sequence_criterion(): Evaluate a sequence initial design
magp_quantitative_criterion(): Evaluate a quantitative Latin hypercube
magp_joint_criterion(): Evaluate a complete quantitative-sequence initial design

Use expected improvement for sequential optimization.

magp_expected_improvement(): Score candidate experiments with expected improvement
magp_next_point(): Select the next quantitative-sequence experiment
magp_bayes_optimize(): Continue Bayesian optimization from completed experiments
magp_bayes_optimize_from_scratch(): Start Bayesian optimization before any experiments have been run

Print concise summaries of fitted models, designs, and searches.

print(<magp2d>): Summarize a fitted two-dimensional MaGP model
print(<magpfull>): Summarize a fitted full-mapping MaGP model
print(<magp_initial_design>): Print a quantitative-sequence initial design
print(<magp_quantitative_design>): Print a quantitative initial design
print(<magp_sequence_design>): Print a sequence initial design
print(<magp_next_point>): Print a MaGP acquisition-search result
print(<magp_bayes_opt>): Print a MaGP Bayesian optimization result