Speedup correlation computation

jecisc · jecisc · commit 98dd08b68928 · 2023-03-15T15:14:33.000+01:00
diff --git a/src/DataFrame-Math/DataPearsonCorrelationMethod.class.st b/src/DataFrame-Math/DataPearsonCorrelationMethod.class.st
@@ -11,8 +11,8 @@ DataPearsonCorrelationMethod class >> between: x and: y [
 	| xDeviation yDeviation |
 	x size = y size ifFalse: [ SizeMismatch signal: 'Correlation can not be calculated for two series of different size' ].
 
-	xDeviation := x replaceNilsWithZeros - x average.
-	yDeviation := y replaceNilsWithZeros - y average.
+	xDeviation := (x values replaceAll: nil with: 0) - x average.
+	yDeviation := (y values replaceAll: nil with: 0) - y average.
 
 	^ (xDeviation * yDeviation) sum / ((xDeviation ** 2) sum * (yDeviation ** 2) sum) sqrt
 ]

Original file line number	Diff line number	Diff line change
`@@ -11,8 +11,8 @@ DataPearsonCorrelationMethod class >> between: x and: y [`
`11`	`11`	`\| xDeviation yDeviation \|`
`12`	`12`	`x size = y size ifFalse: [ SizeMismatch signal: 'Correlation can not be calculated for two series of different size' ].`
`13`	`13`
`14`		`- xDeviation := x replaceNilsWithZeros - x average.`
`15`		`- yDeviation := y replaceNilsWithZeros - y average.`
	`14`	`+ xDeviation := (x values replaceAll: nil with: 0) - x average.`
	`15`	`+ yDeviation := (y values replaceAll: nil with: 0) - y average.`
`16`	`16`
`17`	`17`	`^ (xDeviation * yDeviation) sum / ((xDeviation ** 2) sum * (yDeviation ** 2) sum) sqrt`
`18`	`18`	`]`