It's an Association Rule Mining and Anomaly Detection algorithm which is considered an improvement on Apriori Algorithm for finding frequent itemsets. It works by creating a Frequent Pattern Tree (FP-Tree) for each data item and then mining it to find frequent itemsets.

The Frequent Pattern Tree (FP-Tree) is a tree structure, With null as it's root node, and other nodes represent frequent itemsets and their frequencies .


  • The results are highly explainable and interpretable.
  • It is less resource intensive than Apriori Algorithm, as it only scan the dataset twice and database is stored in a compact version in memory.
  • It is both efficient and scalable for mining long and short frequent patterns


  • The algorithm is complex and can be computationally expensive to build.
  • In large datasets, it may not be possible to fit in the shared memory.
  • It may generate more itemsets which may requires domain knowledge to make it explainable.