mining quantitative association rules in large relational tables（量化关联规则挖掘在很大关系表）.pdfVIP

Mining Quantitative Asso ciation Rules in Large Relational Tables Ramakrishnan Srikant  Rakesh Agrawal IBM Almaden Research Center IBM Almaden Research Center San Jose CA San Jose CA Abstract table has an attribute corresp onding to each item We intro duce the problem of mining asso ciation rules in and a record corresp onding to each transaction The large relational tables containing b oth quantitative and value of an attribute for a given record is if the categorical attributes An example of such an asso ciation item corresp onding to the attribute is present in the might b e of married p eople b etween age and have transaction corresp onding to the record else In at least cars We deal with quantitative attributes by ne the rest of the pap er we refer to this problem as the partitionin g the values of the attribute and then combining Bo olean Asso ciation Rules problem adjacent partitions as necessary We intro duce measures of Relational tables in most business and scientic partial completeness which quantify the information lost due domains have richer attribute typ es Attributes can to partitioning A direct applicatio n of this technique can b e quantitative eg age income or categorical eg generate to o many similar rules We tackle this problem zip co de make of car Bo olean attributes can b e by using a greaterthanexp ectedvalu e interest measure considered a sp ecial case of cat