How to stop the display of objective function model calculation and the graphs when using GP

I am computing a GP at every iteration using a for-loop. The objective function graphs and calculations in the command wind is taking a lot of space. I have about 201 figures. How can I stop the display of objective function calculation in the command window and the graphs.

N = 100;

x1 = linspace(0, 10, N);
x2 = linspace(0, 5, N);

X = [x1; x2];
X = X';

Y = sin(X(:, 1).^2 + X(:, 2).^2);

gpr = fitrgp(X,Y,'KernelFunction','squaredexponential',...
            'OptimizeHyperparameters','auto','HyperparameterOptimizationOptions',...
            struct('AcquisitionFunctionName','expected-improvement-plus'));

|=====================================================================================================|
| Iter | Eval   | Objective:  | Objective   | BestSoFar   | BestSoFar   |        Sigma |  Standardize |
|      | result | log(1+loss) | runtime     | (observed)  | (estim.)    |              |              |
|=====================================================================================================|
|    1 | Best   |     0.34951 |      1.3498 |     0.34951 |     0.34951 |   0.00019035 |         true |
|    2 | Accept |     0.39379 |     0.23818 |     0.34951 |     0.35264 |       1.4136 |        false |
|    3 | Best   |     0.34824 |     0.26209 |     0.34824 |     0.34829 |   0.00055134 |        false |
|    4 | Accept |     0.39379 |      0.1279 |     0.34824 |     0.34826 |       1.0587 |         true |
|    5 | Accept |     0.34824 |     0.41387 |     0.34824 |     0.34824 |   0.00024395 |        false |
|    6 | Accept |     0.34951 |     0.15481 |     0.34824 |     0.34824 |     0.001418 |         true |
|    7 | Accept |     0.34951 |     0.14522 |     0.34824 |     0.34824 |   0.00053162 |         true |
|    8 | Accept |     0.34824 |     0.14395 |     0.34824 |      0.3482 |   0.00010015 |        false |
|    9 | Accept |     0.34824 |     0.12495 |     0.34824 |     0.34821 |    0.0027851 |        false |
|   10 | Accept |     0.34824 |     0.13997 |     0.34824 |     0.34818 |     0.001432 |        false |
|   11 | Accept |     0.34824 |     0.11798 |     0.34824 |      0.3482 |    0.0016966 |        false |
|   12 | Accept |     0.34824 |     0.15227 |     0.34824 |      0.3482 |   0.00016355 |        false |
|   13 | Accept |     0.34824 |     0.12321 |     0.34824 |      0.3482 |   0.00010024 |        false |
|   14 | Accept |     0.34824 |     0.15397 |     0.34824 |     0.34821 |    0.0018175 |        false |
|   15 | Accept |     0.34824 |     0.11096 |     0.34824 |     0.34821 |   0.00034393 |        false |
|   16 | Accept |     0.34824 |     0.11801 |     0.34824 |     0.34822 |   0.00091657 |        false |
|   17 | Accept |     0.34824 |     0.12112 |     0.34824 |     0.34822 |   0.00016423 |        false |
|   18 | Accept |     0.34824 |     0.10819 |     0.34824 |     0.34822 |     0.001893 |        false |
|   19 | Accept |     0.34824 |     0.14795 |     0.34824 |     0.34822 |       0.0001 |        false |
|   20 | Accept |     0.39379 |    0.085699 |     0.34824 |     0.34822 |       6.8859 |         true |
|=====================================================================================================|
| Iter | Eval   | Objective:  | Objective   | BestSoFar   | BestSoFar   |        Sigma |  Standardize |
|      | result | log(1+loss) | runtime     | (observed)  | (estim.)    |              |              |
|=====================================================================================================|
|   21 | Accept |     0.34824 |       0.171 |     0.34824 |     0.34824 |    0.0056218 |        false |
|   22 | Accept |     0.34824 |     0.10411 |     0.34824 |     0.34825 |    0.0037831 |        false |
|   23 | Accept |     0.34824 |     0.10866 |     0.34824 |     0.34821 |    0.0037717 |        false |
|   24 | Accept |     0.34951 |     0.17323 |     0.34824 |     0.34821 |   0.00010003 |         true |
|   25 | Accept |     0.34824 |     0.20197 |     0.34824 |     0.34822 |    0.0037079 |        false |
|   26 | Accept |     0.34824 |     0.11987 |     0.34824 |     0.34822 |   0.00064846 |        false |
|   27 | Accept |     0.39379 |     0.13155 |     0.34824 |     0.34821 |       6.8888 |        false |
|   28 | Best   |      0.3436 |      0.1614 |      0.3436 |     0.34361 |     0.036807 |         true |
|   29 | Best   |      0.3353 |     0.18172 |      0.3353 |     0.33532 |     0.014568 |         true |
|   30 | Accept |     0.33807 |     0.10861 |      0.3353 |     0.33541 |     0.011318 |         true |

__________________________________________________________
Optimization completed.
MaxObjectiveEvaluations of 30 reached.
Total function evaluations: 30
Total elapsed time: 21.1394 seconds
Total objective function evaluation time: 5.8023

Best observed feasible point:
     Sigma      Standardize
    ________    ___________

    0.014568       true    

Observed objective function value = 0.3353
Estimated objective function value = 0.33541
Function evaluation time = 0.18172

Best estimated feasible point (according to models):
     Sigma      Standardize
    ________    ___________

    0.014568       true    

Estimated objective function value = 0.33541
Estimated function evaluation time = 0.15636

[Telema Harry]

@MathWorks Support Team please any solution to my question? My code takes a long time to run and displaying all the results means that I can't use my system effectively when the code is running. I.e. My current window closes when every new iteration.

Answer 1

[Ayush Aniket]

Hi Telema,

As per my understanding you would like to stop the display of objective function model calculation and the graphs. You can use the ‘ShowPlots’ and ‘Verbose’ options available in the ‘HyperparameterOptimizationOptions’ name-value argument of the function “fitgrp”. These options give you the ability to suppress the display of plots and iterative calculation information. You can modify your code as follows:

% Set the options for hyperparameter optimization to suppress display and plots
opts = struct('AcquisitionFunctionName', 'expected-improvement-plus', ...
              'ShowPlots', false, ... % This stops the display of plots
              'Verbose', 0);          % This stops the display in the command window

gpr = fitrgp(X, Y, 'KernelFunction', 'squaredexponential', ...
             'OptimizeHyperparameters', 'auto', ...
             'HyperparameterOptimizationOptions', opts);

To read more about the options and their functionality please refer to the following documentation page:

https://www.mathworks.com/help/stats/fitrgp.html#d126e462488

Hope it helps!