[1] Mining Association Rules between Sets of Items in Large Databases电子书.pdf

Mining Asso ciation Rules b etween Sets of Items in Large Databases Rakesh Agrawal Tomasz Imielinski Arun Swami IBM Almaden Research Center 650 Harry Road, San Jose, CA 95120 Abstract Several organizations have collected massive amounts of such data. These data sets are usually stored We are given a large database of customer transactions. Each transaction consists of items purchased by a customer on tertiary storage and are very slowly migrating to in a visit. We present an ecient algorithm that generates all database systems. One of the main reasons for the signicant asso ciation rules b etween items in the database. limited success of database systems in this area is The algorithm incorp orates buer management and novel that current database systems do not provide necessary estimation and pruning techniques. We also present results functionality for a user interested in taking advantage of applying this algorithm to sales data obtained from a of this information. large retailing company, which shows the eectiveness of the This pap er intro duces the problem of \mining a large algorithm. collection of basket data typ e transactions for asso cia- tion rules b etween sets of items with some minimum 1 Intro duction sp ecied condence, and presents an ecient algori