Hi,
Yes, I understand that this is a complex issue indeed. A good paper that talks about different counting methods for sequential patterns is [1]. I saw this paper while using a popular R package for sequence analysis, TraMineR. In that package, the function
seqefsub allows to search for frequent sequential patterns using different counting methods (
Function documentation). Unfortunately, I tested it with my dataset and it crashed due to the relying underlying algorithm is quite old (I think it is PrefixSpan). That is why I think that the inclusion of more counting methods in SPMF could be of great help for some cases.
Best!
[1] Joshi, Mahesh V., George Karypis, and Vipin Kumar. "A universal formulation of sequential patterns." Proceedings of the KDD’2001 workshop on Temporal Data Mining. 1999.