Bayesian-OverDBC: A Bayesian Density-Based Approach for Modeling Overlapping Clusters

Algorithm 1

Graphic Bayesian model for overlap clustering.

Overlapping Density Based Clustering Algorithm (OverDBC)

Input: Expression Matrix ()

output: Overlap clusters set ()

//phase 1: For each two point and in

  Find the value of Similarity matrix ()

//phase 2: Find_core_list();

//phase 3: For all of core_list do

     = next_cluster

    expandcluster(, , Neighbors);

    add to

//phase 4: Func_bound_over();

Expandcluster ( , , Neighbors)

  = Link list.new ( );

 For each point in neighbors

   If is in Volume()

   Neighbors = neighbors neighbors’

Return

Find_core_list ( )

 Core_list = undefined

 For each in

  If Density() > avg_Density

   Core_list.insert();

 Core_list. Sort( ) base on Closeness Centrality value

 Return Sorted Core_list;

Func_bound_over ( )

 Compute (the maximum number of overlap point)

For all , in C

If  >=  then

  = merge();

 Delete from

Return