Association Rule Mining III: Frequent Pattern Trees
- 26 Downloads
This chapter introduces the FP-growth algorithm for extracting frequent itemsets from a database of transactions. First the database is processed to produce a data structure called a FP-tree, then the tree is processed recursively by constructing a sequence of reduced trees known as conditional FP-trees, from which the frequent itemsets are extracted. The algorithm has the very desirable feature of requiring only two scans through the database.