Mining Frequent Patterns in Object-Oriented Data
Petr Kuba, Lubomír Popelínský · 2004
This paper tackles mining frequent patterns and classification in object-oriented data. An object-oriented model with its rich structures and database systems support is a good choice for storing complex structured data. A new OR-FP algorithm for mining frequent patterns in object-oriented data is introduced and its implementation described. The OR-FP system loads the data from the Oracle object-relational database system and requires only minimum mandatory settings. Propositionalization with OR-FP for two benchmark domains is described. A modification of OR-FP for processing XML data is discussed