[SPARK-11581] Example mllib code in documentation incorrectly computes MSE

• Type: Bug
• Status: Resolved
• Priority: Trivial
• Resolution: Fixed
• Affects Version/s: 1.3.1, 1.4.1, 1.5.1, 1.6.0
• Fix Version/s: 1.4.2, 1.5.3, 1.6.0
• Component/s: Documentation
• Labels: starter
• Target Version/s: 1.4.2, 1.5.3, 1.6.0
• Assignee: M Bharat lal
• Reporter: Brian Webb
• Created: 09/Nov/15 04:03
• Updated: 10/Nov/15 08:50
• Resolved: 09/Nov/15 19:35

Description

The example Java code at the bottom of the mllib-decision-tree web page shows how to compute MSE on the test data. However, there is a bug in the code. The code currently divides by data.count(), but it should instead divide by the count of testData, testData.count().

http://spark.apache.org/docs/latest/mllib-decision-tree.html

Double testMSE =
predictionAndLabel.map(new Function<Tuple2<Double, Double>, Double>() {
@Override
public Double call(Tuple2<Double, Double> pl)

{ Double diff = pl._1() - pl._2(); return diff * diff; }

}).reduce(new Function2<Double, Double, Double>() {
@Override
public Double call(Double a, Double b)

{ return a + b; }

}) / data.count();
System.out.println("Test Mean Squared Error: " + testMSE);

Issue Links

• links to: [Github] Pull Request #9560 (bharatl)