This project is read-only.
Project Description
Mining frequent patterns is an essential task in a lot of fields,
primarly market basket analysis; you can learn which products are
commonly purchased together and how likely a particular product is to be purchased
along with another. For example, you might find that 5% of your customers
have bought pasta, tomatoes and milk together, and that 75% of those
customers that bought pasta and tomatoes also bought milk.
There are two steps in an association algorithm, the first step is a
calculation intensive process that computes the frequent itemsets
(i.e. the sets of distinct products purchased together more
frequently). The second one generates the association rules from all
the frequent itemsets. This step requires much less time than the
first does. We implemented two well known algorithms: Apriori and
FPGrowth: both of them include a lot of optimization designed to
decrease memory usage and processing time.

FPMiner Paper (italian)
AprioriFPGrowth

Last edited Mar 26, 2008 at 11:38 AM by lukadt, version 3