diff --git a/Tutorial/ActualTutorial/VisualizingData/tutorialVisualizingData.html b/Tutorial/ActualTutorial/VisualizingData/tutorialVisualizingData.html index fa8189c..f230c4f 100644 --- a/Tutorial/ActualTutorial/VisualizingData/tutorialVisualizingData.html +++ b/Tutorial/ActualTutorial/VisualizingData/tutorialVisualizingData.html @@ -6,7 +6,7 @@ tutorialVisualizingData

Contents

Tutorial: How to visualize data

close all
+  
Contents
Tutorial: How to visualize data
close all
 
Test Function
The test function as follows:
$$f_{\rm{Test}} = \frac{10x_{1}}{5 + x_{1}}\sin\left( x_{2} \right) +
x_{3}$$
That is, the output is related to the first input variable by a Michaelis Menten curve(steep increase for small values followed by a plateau). The second input variable leads to oscillation and the third input variable leads to an linear monotonically increase. The effect of the first and second input are coupled with each other.
For a better visualization, pair-wise combination of input variables are plotted against the output. Remaining input variable was hold constant at the value of 2.
visualizeTestFunction();
 
   
Initialization
Step 1: Create a new Kriging analysis object. For more details see documentation of tutorialCreatingObject
[KrigingObj]=tutorialCreatingObject(false);
 
@@ -100,21 +100,40 @@
     set(gca,'FontSize',10)
 end
 KrigingObj.setShowBounds(true)
-
   
Screening Plots
Screening plots show interpolation for all pair-wise combinations of input variables. The remaining input variables are set fix to a predefined reference point.
KrigingObj.setReferencePoint([2,2,2]); % Values for remaining input variables
+
   
Contour plots
Contour plots represent the top view of the 3D Interpolation. It projectes the 3D information to 2D and is often easier to interprete. Before a contour plot can be create the 3D information have to be calculated in the same fashion as for the 3D interpolation plots.
NOTE: Settings such as "ShowData" are global and are used also for several other visualization features
combinations = [1,2,3;1,3,2;2,3,1]; % Pair-wise combinations of input variables
+KrigingObj.setShowData(true) % Show provided data points
+
+% Set the inputs for "plotInterpolation_23D"
+dimension = 3;
+showExpectedImprovement = false;
+showContourplot = true;
+
+for iComb = 1:nInputVar
+    KrigingObj.calcInterpolation_3D(indexKrigingObj,...
+               combinations(iComb,1:2),combinations(iComb,3),2)
+    KrigingObj.plotInterpolation_23D(indexKrigingObj,dimension,...
+                                   showExpectedImprovement,showContourplot)
+    grid on
+
+    % Set format appropriate for documentation
+    set(gcf,'Position', [100, 100, 330, 330/4*3]);
+    set(gca,'FontSize',10)
+end
+
   
Screening Plots
Screening plots show interpolation for all pair-wise combinations of input variables. The remaining input variables are set fix to a predefined reference point.
KrigingObj.setReferencePoint([2,2,2]); % Values for remaining input variables
 KrigingObj.calcScreeningAnalysis(indexKrigingObj)
 KrigingObj.plotScreeningAnalysisKrigingInterpolation(indexKrigingObj)
 
 % Set format appropriate for documentation
 set(gcf,'Position', [100, 100, 700, 700/4*3]);
 set(gca,'FontSize',10)
-
 
nD Interpolation plots
The idea of "n-D" is to use several contour plot showing the dependency of the output variable on two input variables while increasing a third one. This can be repeated for all combination of input variables leading to an overall understanding.
KrigingObj.setnPlots(5); % Number of columns in each row
+
 
nD Interpolation plots
The idea of "n-D" is to use several contour plot showing the dependency of the output variable on two input variables while increasing a third one. This can be repeated for all combination of input variables leading to an overall understanding.
KrigingObj.setnPlots(5); % Number of columns in each row
 KrigingObj.calcInterpolation_nD(indexKrigingObj,[1,2,3;1,3,2;2,3,1],[],[]) % no remaining input variables
 KrigingObj.plotInterpolation_nD(indexKrigingObj)
 
 % Set format appropriate for documentation
 set(gcf,'Position', [100, 100, 700, 700/4*3]);
 set(gca,'FontSize',10)
-
 
MakeMovie
The movie analysis allows a dynamic visualization of three factor interaction. In the end, an avi file is produced with a user defined frame rate. In total, n 3D-interpolation plots are generared for the movie. The camera position has to be adjusted at the beginning and is hold fixed for the remaining recording.
Under Windows, the resulting movie can be opened here
fileName = 'movieAnalysis';
+
 
MakeMovie
The movie analysis allows a dynamic visualization of three factor interaction. In the end, an avi file is produced with a user defined frame rate. In total, n 3D-interpolation plots are generared for the movie. The camera position has to be adjusted at the beginning and is hold fixed for the remaining recording.
Under Windows, the resulting movie can be opened here
fileName = 'movieAnalysis';
 KrigingObj.setFrameRate(2)
 KrigingObj.setnStepsBetweenBounds(10)
 KrigingObj.setReferencePoint([2,2,2]); % Values for remaining input variables
@@ -132,7 +151,7 @@
 Step 8 of 10
 Step 9 of 10
 Step 10 of 10
-

Published with MATLAB® R2015b