Performance of negative association rule mining using improved frequent pattern tree

Abstract

Negative Association Rule (NAR) mining is a task of finding rules which contains data items that are negatively correlated. Many algorithms are implemented to discover the NAR; from the offered approach, the frequent pattern growth (FP-Growth) approach is proficient for finding the item sets, from which we can discover NAR. But in the FP-Growth, it finds numerous Conditional FP Trees (CFP-Tree). Frequent Item Set Mining (FISM) algorithm uses an improved FP-Tree for generating NAR without producing CFP-Tree. In this paper, we presented the overview of three different algorithms: Apriori, FP-Growth, and FISM that can be used in the process of discovering NAR. Finally, we analyze the behavior of these algorithms by considering a simple transactional database.