Mining Sequential Patterns-英文文献.pdfVIP

Mining Sequential Patterns Rakesh Agrawal Ramakrishnan Srikant  IBM Almaden Research Center Harry Road San Jose CA Abstract that customers typically rent Star Wars then Em pire Strikes Back and then Return of the Jedi We are given a large database of customer transac Note that these rentals need not b e consecutive Cus tions where each transaction consists of customerid tomers who rent some other videos in b etween also transaction time and the items b ought in the transac supp ort this sequential pattern Elements of a sequen tion We intro duce the problem of mining sequential tial pattern need not b e simple items Fitted Sheet patterns over such databases We present three algo and at sheet and pillow cases followed by com rithms to solve this problem and empirically evalu forter followed by drap es and rues is an example ate their p erformance using synthetic data Two of of a sequential pattern in which the elements are sets the prop osed algorithms AprioriSome and Apriori of items All have comparable p erformance alb eit AprioriSome p erforms a little b etter when the minimum numb er Problem Statement We are given a database D of customers that must supp ort a sequential pattern of customer transactions Each transaction consists is low Scaleup exp eriments show that b oth Apri of the following elds customerid transactiontime oriSome and AprioriAll scale linearly with the num and the items purchased in the transaction No cus b er of customer transactions They also have excel tomer has more than one