Over one hundred FIM algorithms were proposed - the majority claiming to be the most efficient. To clarify this chaos and the contradictions, two FIMI competitions were organized. It not only resulted in some excellent implementations but the community also get better understanding (limits, drawbacks, advantages) of the algorithms.
The separate work of different research groups, however, cannot be so efficient than tightly cooperating groups that interchange their knowledge. The goal of this FIM environment is to support this tight cooperation by providing an environment that includes base library functions, like IO, coding, decoding, ... and some high level desciption of the most important algorithms, such as apriori, eclat, FP-growth together with some very efficient implementations of them. The enviroment also contains many test classes to test the solutions for specific problems.
Flexibility and plugability are achieved by a template-based solution, that does not prevent the compiler form code optimization (like virtual function-based solution), which results in codes that are competitive with the best solutions.
1.4.4