frequent pattern c 数据仓储与数据挖掘 知识介绍.pptVIP

关联规则 Write-based、DFS算法 VIPER Implement DFS with bitmap compression of data in vertical format FP-Growth算法 Implement DFS using a tree structure in horizontal format FP-Growth based sub-graph mining ;Comparison of Data Layouts;Vertical format for write-based FPM;Vertical format for write-based FPM;关联规则 Write-based、DFS算法 VIPER Implement DFS with bitmap compression of data in vertical format FP-Growth算法 Implement DFS using a tree structure in horizontal format FP-Growth based sub-graph mining ;第一次扫描数据库： 类似于Apriori算法，找出频繁的1-itemset和他们的计 数值， 将频繁项目按频度降序排列 2. 第二次扫描数据库： 构造fp-tree （频度从大到小） 挖掘该树（频度从小到大） ? 注意：数据库大时，fp-tree可能在内存中装不下，需要 采取partition方法。;Construct FP-tree from a Transaction Database;Find Patterns Having P From P-conditional Database;From Conditional Pattern-bases to Conditional FP-trees ;Benefits of the FP-tree Structure;Partition Patterns and Databases;FP-Growth vs. Apriori: Scalability With the Support Threshold;VIPER vs. FP-Growth: Comparision;VIPER vs. FP-Growth: Comparision;Why would write-based algorithms be faster ?;Why would write-based algorithms be faster ?;However......;;No candidates ?;So far, we covered;关联规则 Write-based、DFS算法 VIPER Implement DFS with bitmap compression of data in vertical format FP-Growth算法 Implement DFS using a tree structure in horizontal format FP-Growth based sub-graph mining ;Sub-Graph Mining---pattern growth method;Mofa (ICDM’02);gSpan (Yan Han, ICDM’02);gSpan (cont.);Right-Most Extension;Right-Most Extension (cont.);Avoid Duplicate Extension;Right-most Extension;Search Space in gSpan;DFS Code;DFS Code (cont.);Reduce the Number of G1, … Gn;Right-Most Extension (cont.);So far, we covered;References;References