ASCF: Optimization of the Apriori Algorithm Using Spark-Based Cuckoo Filter Structure

Algorithm 2

ASCF phase two.

	INPUT: R = Transactions (lists of items resulting from phase 1), min_sup = Minimum Support, and Frequent Items = RDD of all the frequent items that are stored in Cuckoo filter.
	OUTPUT: K-F-Itemsets = RDD of all the k-frequent itemsets.
	Foreach transaction T in R
	filter(T: len (T ≥ k))
	End Foreach
	mapPartitions (delete non frequent items)
	Do while k ≥ 2
	Foreach transaction T in Pruned transactions
	filter (T: len (T ≥ k))
	End Foreach
	mapPartitions (MakePairOfkItems)
	Foreach list in lists of pairs of k items
	flatmap (list: pair of k items)
	Foreach pair of pairs of k items
	map (pair: (Pair, 1))
	End Foreach
	End Foreach
	ReduceByKey (Pair, Count)
	End ReduceByKey
	K-F-Itemsets = filter (findFrequent (min_sup))
	IF K-F-Itemsets.count() >1 then
	Foreach itemset in K-F-Itemsets
	F = flatmap (itemset: get Items).distinct()
	End Foreach
	Differ = Frequent Items. Subtract (F)
	IF Differ! = [ ]
	Cf.delete (Differ)
	K = k + 1
	Frequent items = F
	Foreach transaction T in Pruned transactions
	filter (T: len(T ≥ k))
	End Foreach
	mapPartitions (delete non frequent items)
	Else k = k + 1
	Else k = 1
	End