The meta-classifier is a filter on the decision tree results. It can be used to fine tune the classification through the usage of dedicated counters.
By setting these counters the number of transitions on the MLC outputs, and consequently the false positives, can be reduced.
