Modified Anti-Monotone Support Pruning on FP Tree for Improved Frequent Pattern Generation

摘要: Pattern discovery plays a very important role in mining interesting frequent patterns from any data sources. Besides, it also includes determining whether the pattern is interesting. The study applied modified anti-monotone support constraint to Frequent Pattern-Growth (FP-Growth) algorithm enhance its generation. alteration after constructing FP tree, thereby traversing and searching nodes of tree for possible not satisfying minimum support. process resulted comparable difference number generated itemsets association rules. second pruning showed more compared those only first pruning. used confidence measure assessing algorithm's performance. result evaluation that two testing, few obtained most measures, having 1 out 6 itemset S40 5 20 S82. On other hand, all measures. Hence, dual removes infrequent weak patterns, leaving strong ones.