Currently those algorithms to mine association rules only pay attention to one aspect of efficiency or accuracy or correlativity respectively, even they ignore mining principal factors among all the correlativity. Thus, there seems a paradox among efficiency, accuracy and correlativity. In order to resolve to this conflict, a novel algorithm based on Probability estimate and principal component analysis is proposed to mine the association rules from database with the high correlativity and the high confidence. Probability estimate reduce the times of database scanning so as to increase efficiency and accuracy, and principal component analysis helps us to know which factors have most influence to event rate so as to distinguish correlativity. Experimental results have demonstrated that our algorithms are efficient accurate and correlativity.
[1]
Hu Jun.
A Fault Diagnosis Method Based on Simulation and Data Mining
,
2007
.
[2]
Petra Perner,et al.
Data Mining - Concepts and Techniques
,
2002,
Künstliche Intell..
[3]
Chris Clifton,et al.
Query flocks: a generalization of association-rule mining
,
1998,
SIGMOD '98.
[4]
Ramakrishnan Srikant,et al.
Mining Association Rules with Item Constraints
,
1997,
KDD.
[5]
Heikki Mannila,et al.
Finding interesting rules from large sets of discovered association rules
,
1994,
CIKM '94.
[6]
浙江大学.
高校应用数学学报. Ser. A = Applied mathematics : a journal of Chinese universities
,
1993
.