基于关联的自适应分类规则挖掘模型的研究-计算机科学与技术专业论文.docxVIP

攘要撼要 攘要 撼要 分炎是数强挖掘中应曩j褥簸多麴强务之～。在CRM(罄户关系蛰理)中， 分类分橱就燕通过分析梯零客户数据库中的数掭，为每个类别作出准确的擒述或 建立分辑摸纛或挖掘趣分类疑划，然瑶蠲这个分类规则对其它客户的记录滋罩予分 类。疆髓分类的磷究多偏重于从冀法的角度进符研究，主要集中在提蕊分类算法 晌准确攫帮效率上，对于鲡秘实现一个完善稳嫩弼系统则没寿绘予太多的关注， 而数据挖掘憝～个与应用结合如眈紧密的一项研究，西j琏：绪台其俸的威焉需求对 数摇挖掘进褥磷究是十分必要瓣。 本文提出了一个基于关联的自逶成分类瀵燹|j挖黼摸鹜，主骚研究葶疆探讨了浚 模登在预链疆、多麓分黉囊舞|j豹挖瓣、簿法熊W扩溪靛、效率帮输入参数豹自适 应等方蕊的憨慧、技术和方法。在～般静蕊子关联靛分类方法巾，帮怒辎爝檬灌 关联麓瓣挖攘冀法(翔npnori算法)挖糕懑露关斡美联窳粼，本文箍密了巅嗣 不需爱产生候选集鹣簇繁集撩籀雾滚FP．Growth算法卷产黛频繁集，大大降低了 辩阕鞠窆闻笺杂度。弱瓣对渡雾法遴毒亍了改造，使冀熬逶窿关联分类瓣粼懿挖掘， 遥避决策属性中覆馁僖的最大支持发max sup及时消豫无蘸途瓣搜索姆径，减少 7嚣衾频繁簇黪产黛，褥嵩了挖掘瓣骞效蕊。筠辩，搜输入参数疑叁邋应遴雩亍调 整，释低了搦声输入拽零参数蠡{簦难度，疑离了熬令搂羹jc|分类瓣羹|l捡搦瓣簧糁榷、 准确毪、效零帮霉扩疑瞧。爱嚣对辘爨筑甏港行趸佘羧分瓣，测除嚣佘熬浚窭溪 巅，馊耩爨簸剃姥雯翱灞足耀声霉蘩辫覆其鸯疆学拣。 荚键溜：数器羧撬分爨怒粼魏攮基予关联壹逶畿 AbstractABSTRACT Abstract ABSTRACT Classification is one of也e most commonly used tasks in data mining．In CRM (Customer Relationship Management)，classification analysis make exact description or build analysis model or mine classification rules for every category throu蜘analyzing the data in sample customer databases．Then other customer records arc classified by the classitication rules．Recently,most of analysis about classification is laid particular stress on arithmetic and fasten on enhance the nicety and efficiency of classification arithmetic．A few of investigators pay attention to that how to achieve a perfect application system．However,data mining is an analysis that integrated with application closely．It is very necessary that studying on data mining link with requirement of material applications． The paper proposes a model ofassociation based self-adaptive mining classification rules，and makes a deep research and discussion on ideas，technologies and methods about the pre-processing，the mining on multilevel classification rules，the extensible and efficiency of arithmetic and the self-adaptive of import parameters etc．In generic assoeiati妊based classifie：ation methods,using standard association rules mining arithmetic(for example：Apriori arithmetic)to mining relational association rules．The paper pr