I am interested in latest algorithms of multilevel association rule mining which are scalable in nature. How can rule pruning be done during mining process so that subsequent passes of algorithm work efficiently?
If multilevel association rule mining is not dealt with directly in formal concept analysis based work on association rules, there are some clear links that show the way. First, association rules and the attribute logic of formal concept analysis are essentially the same thing (though FCA doesn't worry about frequencies and support). Second, FCA is used in deriving ontologies and so there are ways to deal effectively with the "multilevel" business. And, third, the mathematics of apposition/supposition make it possible to combine the results of computation over properly formed subsets of the data.
- Badges
- Science topic
- Similar topics
- Computer Science
- Data Mining