Discovering Significant Relaxed Order-Preserving Submatrices

author: Qiong Fang, Department of Computer Science and Engineering, The Hong Kong University of Science and Technology
published: Oct. 1, 2010,   recorded: July 2010,   views: 3259


Related Open Educational Resources

Related content

Report a problem or upload files

If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.
Lecture popularity: You need to login to cast your vote.


Mining order-preserving submatrix (OPSM) patterns has received much attention from researchers, since in many sci entific applications, such as those involving gene expression data, it is natural to express the data in a matrix and also important to find the order-preserving submatrix patterns. However, most current work assumes the noise-free OPSM model and thus is not practical in many real situations when sample contamination exists. In this paper, we propose a relaxed OPSM model called ROPSM. The ROPSM model supports mining more reasonable noise-corrupted OPSM patterns than another well-known model called AOPC (approximate order-preserving cluster). While OPSM mining is known to be an NP-hard problem, mining ROPSM patterns is even a harder problem. We propose a novel method called ROPSM-Growth to mine ROPSM patterns. Specifically, two pattern growing strategies, such as column-centric strategy and row-centric strategy, are presented, which are effective to grow the seed OPSMs into significant ROPSMs. An effective median-rank based method is also developed to discover the underlying true order of conditions involved in an ROPSM pattern. Our experiments on a biological dataset show that the ROPSM model better captures the characteristics of noise in gene expression data matrix compared to the AOPC model. Importantly, we find that our approach is able to detect more quality biologically significant patterns with comparable efficiency with the counterparts of AOPC. Specifically, at least 26.6% (75 out of 282) of the patterns mined by our approach are strongly associated with more than 10 gene categories (high biological significance), which is 3 times better than that obtained from using the AOPC approach.

See Also:

Download slides icon Download slides: kdd2010_fang_dsr_01.ppt (1.8┬áMB)

Help icon Streaming Video Help

Link this page

Would you like to put a link to this lecture on your homepage?
Go ahead! Copy the HTML snippet !

Write your own review or comment:

make sure you have javascript enabled or clear this field: