Fast Borderline Mining
Year:
2007
Researcher(s):
Dustin Baumgartner, Ryan Millikin
Institution:
Ohio Northern University
Discipline:
Computer Science
In this paper, we present a modification to the AprioriBL algorithm, which is an extension to a well-known Association Mining algorithm, Apriori. AprioriBL targets the borderline cases of frequent itemsets; however, it performs poorly. Our new algorithm, AprioriBLT, considers only the borderline cases for generating itemsets. This increases performance at the cost of accuracy. A comparison is made between AprioriBL and AprioriBLT, and the efficacy of AprioriBLT is discussed.
ONU Student Research Colloquium