File heuristic_precision.hpp¶
-
namespace seco
-
class PrecisionConfig : public seco::IHeuristicConfig¶
- #include <heuristic_precision.hpp>
Allows to configure a heuristic that measures the fraction of correctly predicted labels among all labels that are covered by a rule.
This heuristic is equivalent to RIPPER’s pruning heuristic (“Fast Effective Rule Induction”, Cohen 1995, see https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=6c34b7b0441bff66cce2418d36acfd9776ad7bd2). A proof is provided in the paper “Roc ‘n’ Rule Learning — Towards a Better Understanding of Covering Algorithms”, Fürnkranz, Flach 2005 (see https://link.springer.com/content/pdf/10.1007/s10994-005-5011-x.pdf).
Public Functions
-
virtual std::unique_ptr<IHeuristicFactory> createHeuristicFactory() const override¶
Creates and returns a new object of type
IHeuristicFactoryaccording to the specified configuration.- Returns:
An unique pointer to an object of type
IHeuristicFactorythat has been created
-
virtual std::unique_ptr<IHeuristicFactory> createHeuristicFactory() const override¶
-
class PrecisionConfig : public seco::IHeuristicConfig¶