Optimizers

Optimization Methods

As you may have already seen in the demos, the optimizer is initialized in the following way:

GP.setOptimizer(method, num_restarts=None, min_threshold=None, meanRange=None, covRange=None, likRange=None)

This method is used to sepecify optimization configuration. By default, gp uses a single run “minimize”.

Parameters:

• method –

  Optimization methods. Possible values are:

  “Minimize” -> minimize by Carl Rasmussen (python implementation of “minimize” in GPML)

  “CG” -> conjugent gradient

  “BFGS” -> quasi-Newton method of Broyden, Fletcher, Goldfarb, and Shanno (BFGS)

  “SCG” -> scaled conjugent gradient (faster than CG)

• num_restarts – Set if you want to run mulitiple times of optimization with different initial guess. It specifys the maximum number of runs/restarts/trials.
• min_threshold – Set if you want to run mulitiple times of optimization with different initial guess. It specifys the threshold of objective function value. Stop optimization when this value is reached.
• meanRange – The range of initial guess for mean hyperparameters. e.g. meanRange = [(-2,2), (-5,5), (0,1)]. Each tuple specifys the range (low, high) of this hyperparameter, This is only the range of initial guess, during optimization process, optimal hyperparameters may go out of this range. (-5,5) for each hyperparameter by default.
• covRange – The range of initial guess for kernel hyperparameters. Usage see meanRange
• likRange – The range of initial guess for likelihood hyperparameters. Usage see meanRange