In November, I released the version 2.64 of SPMF with 5 new algorithms and 2 new tools, among other improvements, but I still had several other pattern mining algorithms waiting to be released after more tests.
Thus, today, I make a new release of SPMF for December 2025, called version 2.64b. This release has several new algorithms:
- The HMP-SA algorithm for discovering compressing itemsets in a transaction database using a simulated annealing approach (Chen, E. et al., 2026).
- The HMP-HC algorithm for discovering compressing itemsets in a transaction database using a hill-climbing approach (Chen, E. et al., 2026).
- the GENMAX algorithm for mining frequent maximal itemsets from a transaction database (Gouda et al., 2005)
- the DIC algorithm for mining frequent itemsets in a transaction database using dynamic itemset counting (Brin et al. 1997)
- the Talky-G and Talky-G-Diffset algorithms for mining frequent generator itemsets in a transaction database (Szathmary, 2009)
- the iMEFIM algorithm for high utility itemset mining, a variation of EFIM that can be faster but can consume more memory (Nguyen et al., 2019)
- the PUCPMiner algorithm for high utility itemset mining, a variation of FHM with a potentially tighter upper bound (Patel et al., 2022
Moreover, there is a new Pattern Diff Analyzer tool that allows to compare two pattern files to find contrast patterns. This tool works letting the user first select two files containing patterns:
And then we can compute the differences between these files using different contrast methods. For example, we can identify all patterns that are in the first file but not in the second, among multiple other options for identifying contrast patterns:
That is all I wanted to share about this new version! Hope that you will enjoy this new release of the SPMF pattern mining software. So far we have released 18 new algorithms this year! And more features are coming soon…
