TY - JOUR
T1 - Adapting the right measures for pattern discovery
T2 - A unified view
AU - Wu, Junjie
AU - Zhu, Shiwei
AU - Xiong, Hui
AU - Chen, Jian
AU - Zhu, Jianming
N1 - Funding Information:
Manuscript received July 6, 2011; revised October 26, 2011 and January 11, 2012; accepted January 26, 2012. Date of publication March 9, 2012; date of current version July 13, 2012. The work of J. Wu was supported in part by the National Natural Science Foundation of China (NSFC) under Grants 71171007, 70901002, 71031001, 70890080, and 90924020 and in part by the Doctoral Fund of Ministry of Education of China under Grant 20091102120014. The work of H. Xiong was supported in part by NSFC under Grant 71028002 and in part by the National Science Foundation under Grants CCF-1018151 and IIP-1069258. The work of J. Chen was supported in part by NSFC under Grant 70890082. The work of J. Zhu was supported in part by NSFC under Grant 60970143. This paper was recommended by Associate Editor M. S. Obaidat.
PY - 2012
Y1 - 2012
N2 - This paper presents a unified view of interestingness measures for interesting pattern discovery. Specifically, we first provide three necessary conditions for interestingness measures being used for association pattern discovery. Then, we reveal one desirable property for interestingness measures: the support-ascending conditional antimonotone property (SA-CAMP). Along this line, we prove that the measures possessing SA-CAMP are suitable for pattern discovery if the itemset-traversal structure is defined by a support-ascending set enumeration tree. In addition, we provide a thorough study on the family of the generalized mean (GM) measure and show their appealing properties, which are exploited for developing the GMiner algorithm for finding interesting association patterns. Finally, experimental results show that GMiner can efficiently identify interesting patterns based on SA-CAMP of the GM measure, even at an extremely low level of support.
AB - This paper presents a unified view of interestingness measures for interesting pattern discovery. Specifically, we first provide three necessary conditions for interestingness measures being used for association pattern discovery. Then, we reveal one desirable property for interestingness measures: the support-ascending conditional antimonotone property (SA-CAMP). Along this line, we prove that the measures possessing SA-CAMP are suitable for pattern discovery if the itemset-traversal structure is defined by a support-ascending set enumeration tree. In addition, we provide a thorough study on the family of the generalized mean (GM) measure and show their appealing properties, which are exploited for developing the GMiner algorithm for finding interesting association patterns. Finally, experimental results show that GMiner can efficiently identify interesting patterns based on SA-CAMP of the GM measure, even at an extremely low level of support.
KW - Conditional antimonotone property (AMP)
KW - correlation computation
KW - generalized mean (GM)
KW - interestingness measure
KW - set enumeration tree (SET)
UR - http://www.scopus.com/inward/record.url?scp=84864145149&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84864145149&partnerID=8YFLogxK
U2 - 10.1109/TSMCB.2012.2188283
DO - 10.1109/TSMCB.2012.2188283
M3 - Article
AN - SCOPUS:84864145149
VL - 42
SP - 1203
EP - 1214
JO - IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
JF - IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
SN - 1083-4419
IS - 4
M1 - 6166904
ER -