[MATH-436] KMeansPlusPlusClusterer Exception when clusters>variables

• Type: Wish
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 3.0
• Fix Version/s: 3.0
• Assignee: Unassigned
• Reporter: Erik van Ingen
• Created: 11/Nov/10 08:24
• Updated: 24/Mar/12 16:16
• Resolved: 23/Mar/11 13:20

Description

It would be nice when KMeansPlusPlusClusterer generates an exception when the number of cluster is larger than the number of variables.

KMeansPlusPlusClustererTest:
/**

• 2 variables cannot be clustered into 3 clusters.
  */
  @Test
  public void testPerformClusterAnalysisToManyClusters() {
  KMeansPlusPlusClusterer<EuclideanIntegerPoint> transformer = new KMeansPlusPlusClusterer<EuclideanIntegerPoint>(
  new Random(1746432956321l));
  EuclideanIntegerPoint[] points = new EuclideanIntegerPoint[]
  Unknown macro: { new EuclideanIntegerPoint(new int[] { 1959, 325100 }), new EuclideanIntegerPoint(new int[] { 1960, 373200 }), }

  ;
  try

  { transformer.cluster(Arrays.asList(points), 3, 1); Assert.fail(); }

  catch (Exception e)

  { // 2 variables cannot be clustered into 3 clusters. }

  }