Active Semisupervised Clustering Algorithm with Label Propagation for Imbalanced and Multidensity Datasets

Algorithm 2

Semisupervised clustering algorithm with label propagation.

(1) Input the value of k.

(2) 𝑆𝑒𝑙𝑒𝑐𝑡𝐷𝑎𝑡𝑎𝑃𝑜𝑖𝑛𝑡𝐷,𝑝.

(3) Suppose that the number of different labels in 𝐷_𝑙 is p.

(4) Merge 𝑇₁,𝑇₂,…,𝑇_𝑚 into 𝐶₁,𝐶₂,…,𝐶_𝑝 according

  to labels of data in 𝐷_𝑙.

(5) Foreach cluster 𝐶 in 𝐶₁,𝐶₂,…,𝐶_𝑝  do

(6) Foreach data point x in C  do

(7)  Compute the KNN(𝑥,𝐶)

(8)   End Foreach

(9)  Compute 𝑘_𝑎𝑣𝑔𝑑𝑖𝑠(𝐶)s

(10) End Foreach

(11) Foreach cluster 𝐶 in 𝐶₁,𝐶₂,…,𝐶_𝑝  do

(12)   𝐸𝑥𝑝𝑒𝑛𝑑(𝐶,𝑘_𝑎𝑣𝑔𝑑𝑖𝑠(𝐶),𝐷_𝑙)

(13) End Foreach

(14) Label the rest unlabeled data according to KNN rule.

(15) Output the clustering results.