Dunn Index
canopy 1、向量序列化(自定义VectorTo.java) 2、执行canopy ./mahout canopy -i /bs/xx -o /output -dm org.apache.mahout.common.distance.EculideanDistanceMeasure -t1 3.0 -t2 1.5 --clustering 簇的集合列表 ./mahout seqdumper -i /bs/can/clusteredPoints/part-m-00000 Key: 0: Value: wt: 1.0 distance: 0.74535599249993 vec: [1:1.000] Key: 0: Value: wt: 1.0 distance: 0.74535599249993 vec: [0:1.000] Key: 0: Value: wt: 1.0 distance: 0.4714045207910317 vec: [] Key: 1: Value: wt: 1.0 distance: 0.745355992499931 vec: [5.000, 6.000] Key: 1: Value: wt: 1.0 distance: 0.745355992499931 vec: [6.000, 5.000] Key: 1: Value: wt: 1.0 distance: 0.4714045207910384 vec: [5.000, 5.000] Key: 2: Value: wt: 1.0 distance: 0.9428090415820768 vec: [10.000, 2.000] Key: 2: Value: wt: 1.0 distance: 0.4714045207910322 vec: [11.000, 3.000] Key: 2: Value: wt: 1.0 distance: 0.4714045207910322 vec: [11.000, 3.000] 簇的中心点 ./mahout seqdumper -i /bs/can/clusters-0-final/part-r-00000 Key: C-0: Value: org.apache.mahout.clustering.iterator.ClusterWritable@200957cd Key: C-1: Value: org.apache.mahout.clustering.iterator.ClusterWritable@200957cd Key: C-2: Value: org.apache.mahout.clustering.iterator.ClusterWritable@200957cd 自定义类VectorParse解析part-r-00000 C-0 {0:0.3333333333333333,1:0.3333333333333333} C-1 {0:5.333333333333333,1:5.333333333333333} C-2 {0:10.666666666666666,1:2.6666666666666665}