[Issue 81] 2 PCA implementations that give same results but different from Python scikit-learn implementation

src/Math-Tests-PrincipalComponentAnalysis/PMPrincipalComponentAnalyserTest.class.st

@@ -1,3 +1,9 @@
+"
+The tests in this class compare the output from PolyMath with: 
+1) an example from Scikit-Learn's documentation; and
+2) the PCA Tutorial by Lindsay Smith (see http://www.cs.otago.ac.nz/cosc453/student_tutorials/principal_components.pdf). The input data used
+are the mean-centred data in section 3.1, step 2.
+"
 Class {
 	#name : #PMPrincipalComponentAnalyserTest,
 	#superclass : #TestCase,
@@ -10,6 +16,41 @@ Class {
 	#category : #'Math-Tests-PrincipalComponentAnalysis'
 }
 
+{ #category : #'pca tutorial' }
+PMPrincipalComponentAnalyserTest >> testJacobiBasedTransformWithMeanCentredMeasurements [
+	"As acceptance tests we use the worked example of the PCA tutorial of Lindsay Smith"
+	| m pca transformedData expected |
+	m := PMMatrix rows: #(
+			#(0.69 0.49) 
+			#(-1.31 -1.21) 
+			#(0.39 0.99) 
+			#(0.09  0.29) 
+			#(1.29 1.09) 
+			#(0.49 0.79) 
+			#(0.19 -0.31) 
+			#(-0.81 -0.81) 
+			#(-0.31 -0.31) 
+			#(-0.71 -1.01)
+		  ).
+	pca := PMPrincipalComponentAnalyserJacobiTransformation new componentsNumber: 2.
+
+	pca fit: m.	
+	transformedData := pca transform: m.
+
+	expected := PMMatrix rows: #(
+						#(-0.827970186 -0.175115307) 
+						#(1.77758033 0.142857227) 
+						#(-0.992197494 0.384374989) 
+						#(-0.274210416 0.130417207) 
+						#(-1.67580142 -0.209498461) 
+						#(-0.912949103 0.175282444) 
+						#(0.0991094375 -0.349824698) 
+						#(1.14457216 0.0464172582) 
+						#(0.438046137 0.0177646297) 
+						#(1.22382056 -0.162675287)).
+	self assert: (transformedData abs) closeTo: expected abs.
+]
+
 { #category : #tests }
 PMPrincipalComponentAnalyserTest >> testPCAwithPCAandJacobiTransformationReturnSame [
 	| m pca1 pca2 |
@@ -22,6 +63,42 @@ PMPrincipalComponentAnalyserTest >> testPCAwithPCAandJacobiTransformationReturnS
 	self assert: pca1 transformMatrix abs closeTo: pca2 transformMatrix abs
 ]
 
+{ #category : #'pca tutorial' }
+PMPrincipalComponentAnalyserTest >> testSVDBasedTransformWithMeanCentredMeasurements [
+	"As acceptance tests we use the worked example of the PCA tutorial of Lindsay Smith"
+
+	| m pca transformedData expected |
+	m := PMMatrix rows: #(
+			#(0.69 0.49) 
+			#(-1.31 -1.21) 
+			#(0.39 0.99) 
+			#(0.09  0.29) 
+			#(1.29 1.09) 
+			#(0.49 0.79) 
+			#(0.19 -0.31) 
+			#(-0.81 -0.81) 
+			#(-0.31 -0.31) 
+			#(-0.71 -1.01)
+		  ).
+	pca := PMPrincipalComponentAnalyserSVD new componentsNumber: 2.
+
+	pca fit: m.	
+	transformedData := pca transform: m.
+
+	expected := PMMatrix rows: #(
+						#(-0.827970186 -0.175115307) 
+						#(1.77758033 0.142857227) 
+						#(-0.992197494 0.384374989) 
+						#(-0.274210416 0.130417207) 
+						#(-1.67580142 -0.209498461) 
+						#(-0.912949103 0.175282444) 
+						#(0.0991094375 -0.349824698) 
+						#(1.14457216 0.0464172582) 
+						#(0.438046137 0.0177646297) 
+						#(1.22382056 -0.162675287)).
+	self assert: (transformedData abs) closeTo: expected abs.
+]
+
 { #category : #tests }
 PMPrincipalComponentAnalyserTest >> testTransformMatrixWithJacobiTransformation [
 	| m pca expected |