Introduction to Data Mining

Lesson 6.6.3.3

Reducing Transactions

	In each scan, remove all transactions which do not contain any frequent k-itemsets. This will reduce the number of transactions to be scanned in the next scans.
	Remove items from transactions, which are not members of any candidate set. Examples:

If 12, 24, 14 are the only candidate sets from the transaction D_i = 1234, then we can remove itemset 3 from the transaction.

Now if 12, 24 are the only candidate sets from the transaction D_i = 1234, then we can remove the entire transaction from the next round, as there cannot be any itemsets of size = 3.

Such reductions in data size lead to lesser reading and processing times, but usually have higher writing times.

<<< Previous Up Next >>>

	If 12, 24, 14 are the only candidate sets from the transaction D_i = 1234, then we can remove itemset 3 from the transaction.
	Now if 12, 24 are the only candidate sets from the transaction D_i = 1234, then we can remove the entire transaction from the next round, as there cannot be any itemsets of size = 3.