ppt课件-towards accurateand efficient classification a frequentand（frequentand向精确高效的分类）.pptVIP

Apriori Approach Join two size-k patterns to a size-(k+1) pattern Itemset: {a,b,c} + {a,b,d} ? {a,b,c,d} Graph: Pattern Growth Approach Depth-first search, grow a size-k pattern to size-(k+1) one by adding one element Frequent subgraph mining Vertical Data Approach Major operation: transaction list intersection Mining High Dimensional Data High dimensional data Microarray data with 10,000 – 100,000 columns Row enumeration rather than column enumeration CARPENTER [Pan et al., KDD’03] COBBLER [Pan et al., SSDBM’04] TD-Close [Liu et al., SDM’06] Mining Colossal Patterns[Zhu et al., ICD