Performance analysis of frequent itemset finding techniques using sparse datasets

Techniques for efficient mining of frequent patterns have been studied extensively by many researchers. However, the previously proposed techniques still encounter some performance bottlenecks when mining databases with different data characteristics such as, dense vs. sparse, long vs. short patterns, memory-based vs. disk-based, etc. In this study, explored the unifying feature among the internal working of various mining techniques such as, Two-Fold Cross Validation Model, Improved Apriori, CBT-fi and Semi-Apriori. Extensive experiments had been carried out and compared using the different techniques, a tentative result reveal that Semi-Apriori outperforms in terms of execution time using hepatitis and wine datasets.