Research Article

A Novel Convex Clustering Method for High-Dimensional Data Using Semiproximal ADMM

Algorithm 1

sPADMM for solving (16).
(1)Initialization Let and . Given a start point , for
(2)repeat
(3) Step 1. Update according to (25)
(4) Step 2. Update according to (28), (29) and (30)
(5) Step 3. Update the Lagrangian multipliers by (21)
(6)until Stopping criterion is satisfied.