[Reading] Lecture 14 - Mining Frequent Patterns without Candidate Generation: A Frequent-Pattern Tree Approach

This paper proposed a efficient algorithm for finding frequency item sets. It utilized a data structure called FP tree to do this job. There are two advantages of this algorithm. First, the FP tree can very compactly represent the data without any information loss. That's because it didn't use any approximation. The solution will be exactly as same as the right answer. Second, the query is efficient. You only have to scan the database once, which is better than those apriori-based method. A deficincy of this work is the memory problem.

Reference:
"Mining Frequent Patterns without Candidate Generation: A Frequent-Pattern Tree Approach," Han, Data Mining and Knowledge Discovery, 2004.