Class strategy::BayesianOpt

Bayesian Global Optimization stategy searching implementation. The BayesianOpt is a wrapper class for GPP library https://github.com/wujian16/Cornell-MOE The paper "A Tutorial on Bayesian Optimization" is a good introduction of bayesian optimization.

Inheritance
strategy::BaseOpt
strategy::BayesianOpt
Inherited Members

Constructors

BayesianOpt()

Declaration
strategy::BayesianOpt::BayesianOpt()

BayesianOpt()

Declaration
strategy::BayesianOpt::BayesianOpt(const BayesianOpt&)=delete

Methods

operator=()

Declaration
BayesianOpt&strategy::BayesianOpt::operator=(const BayesianOpt&)=delete

configure()

Declaration
void strategy::BayesianOpt::configure(const utils::Json&params, int thread_count)

init()

Declaration
void strategy::BayesianOpt::init(size_t dimensions, uint32_t seed, size_t reserved_samples=128)

init()

Declaration
void strategy::BayesianOpt::init(size_t dimensions, uint32_t seed, const GDParameters&model_parameters, const GDParameters&search_parameters, size_t reserved_samples=128)

recommend_parameter_values()

Declaration
bool strategy::BayesianOpt::recommend_parameter_values(std::vector<double>&parameters_new) override

add_new_sample()

Declaration
void strategy::BayesianOpt::add_new_sample(std::vector<double>&parameters, double objective) override

set_ranges()

Declaration
void strategy::BayesianOpt::set_ranges(const std::vector<std::pair<double, double>>&ranges)

num_of_saved_samples()

Declaration
size_t strategy::BayesianOpt::num_of_saved_samples() const

get_sample()

Declaration
double strategy::BayesianOpt::get_sample(int indx, std::vector<double>&sample_point) const

get_reserved_samples()

Declaration
uint32_t strategy::BayesianOpt::get_reserved_samples()

copyout_winner()

Declaration
void strategy::BayesianOpt::copyout_winner(std::vector<double>&parameters) const

log_hyper_parameters()

Declaration
void strategy::BayesianOpt::log_hyper_parameters() const