使用CUES算法进行聚类

Question

我正在使用CUES算法执行文档聚类。根据簇距离，必须合并两个簇。当合并两个簇Ci和Cj时，Ci用Ci U Cj更新并且移除Cj并更新簇的索引。

我在Java中使用Set Class初始化集群。但是，我无法更新特定索引的群集。如果还有其他可以使用的课程，请帮助我。请在下面找到算法。谢谢。

 Input: a) A set of N clusters, C = {C1,C2, ...,CN} and noc = |C|, number of clusters.
 b) Ci = {di} ∀i ∈ N, where di is the ith document of the data set.
 c) A similarity matrix Sim[i][j] = cluster dis(Ci,Cj), ∀i, j ∈ [1,N].
 Steps of the Algorithm:
 1: X ← 0, Y ← 0
 2: while noc > 1 and X ≥ 0 and Y ≥ 0 do
 3:      min dist ← N
 4:      X ← −1, Y ← −1
 5:      for i = 1 to noc − 1 do
 6:         for j = i + 1 to noc do
 7:             if min dist ≥ cluster dis(Ci,Cj) and cluster dis(Ci,Cj) ≥ 0 then
 8:                   min dist ← cluster dis(Ci,Cj)
 9:                      X ← i, Y ← j
10:                 end if
11:             end for
12:           end for
13:          if X ≥ 0 and Y ≥ 0 then
14:                 CX ← CX ∪ CY
15:                 Sim ← merge(Sim, i, j)
16:               noc ← noc − 1
17:         end if
18: end while
19: return C