Consensus Kernel -Means Clustering for Incomplete Multiview Data

Algorithm 1

Consensus kernel -means clustering for incomplete multiview clustering.

Input:

      Incomplete multi-view data:

      Indicator matrix:

      Number of clusters:

      Balance parameter:

      Stopping gap:

Output:

      The consistent decision:

      Filled kernels:

  (1)  Initialize filled kernels with zero-filling

 (2)  Initialize by performing kernel- means on

      initialized filled kernels

 (3)  Initialize by Eq. (10)

 (4)  repeat

 (5)   Update by solving Eq. (11)

 (6)   Update by solving Eq. (8)

 (7)   Update by solving Eq. (10)

 (8)  until Objective difference smaller than

 (9)  return,