利用标准SQL查询挖掘多值型关联规则及其评价.pdfVIP

第39卷第3期 计 算机研 究与发 展 V0_l 39．N0．3 2002年3月 JOURNAI OF COMPUTER RESEARCH AND DEVEI OPMENT Mar．2002 利用标准SQL查询挖掘多值型关联 规则及其评价 杨炳儒① 孙海洪① 熊范纶② (北京科技太学信息32程学院 北京 100083) 0(中国科学院合肥智能所 台肥 230031) (bryang@public fhnet．1211．rtet) 摘 要 关联规则是数据挖掘中一种重要的模式，目前对布尔型关联规则的挖掘算法研究得比较多．利用标准 SQL功能强、易于开发、运行速度快、安全可靠性强等优点，来实现多值型关联规则的挖掘具有比较好的性能．首先 利用语言场理论对连续属性进行离散化，然后利用sqI 对Aprior[算法进行改进来实现关联规则的挖掘．另外利 用主观Bayes方法中的LS充分性因子对挖掘出来的假设规则进行评价．该算法应用于庐江虫害数据库，结果显示 它具有 速、有效、易开发、适用范围广等优点． 关键词 多值型关联规则，SQL语言，评价 中图法分类号 TP3¨ MINING QUANTITATIVE ASSOCIATION RULES WITH STANDARD SQL QUERIES AND IT’S EVALUAT10N YANG Bing—Ru~．SUN Hai Hong~．and XIONG Fan Lun@ (School。j Information Engineering，Bering Univer~,$a,of S~enee t2nd TechnoZogy B∞ ng 100083) H e硒 Institute ofIntell gen￡Machbws，ChineseAcademyof Sconces，Hefm230031 Abstract Association rule is an important model in data mining．Currently，more research work is done on the algorithm of mining boolean association rules than on the algorithm of quantitatire association rules．With standard SQL，which has the advantages of powerful function，easy development．quick running speed，and srlong reliability etc，tO implement the quantitative association rules mining can result in good performance．Firstly，quantitative attribute needs to be dispersed with language field theory．Then association rules are mined by improving Apriori algorithm with SQL．A．dditionly，the association rules are evaluated with the sufficiency gene LS of subjectivity Bayes reasoning．This algorithm has been applied to the Lujiang Insect Pests Database，and