Efficient Block Bi-clustering by Alternating Semidefinite Programming Relaxation

TMLR Paper6971 Authors

Abstract: The bi-clustering problem is one of the most common problems in data mining. In this paper, we solve the block bi-clustering problem by using the semidefinite programming (SDP) relaxation alternately for clustering rows and columns of the data matrix. Theoretically, in common noisy cases, our algorithm can accurately identify the checkerboard pattern; if there is no noise in the data matrix, we establish an exact recovery for the checkerboard pattern. In both simulated and real data experiments, we show that our algorithm performs comparably or better than other bi-clustering methods in terms of both accuracy and efficiency.

Submission Type: Long submission (more than 12 pages of main content)

Changes Since Last Submission: N/A

Assigned Action Editor: ~Patrick_Flaherty1

Submission Number: 6971