This paper presents some empirical results on simplification methods of decision trees induced from data. We observe that those methods exploiting an independent pruning set do not perform uniformly better than the others. Furthermore, a clear definition of bias towards overpruning and underpruning is exploited in order to interpret empirical data concerning the size of the simplified trees.
Simplifying decision trees by pruning and grafting: New results
ESPOSITO, Floriana;SEMERARO, Giovanni
1995-01-01
Abstract
This paper presents some empirical results on simplification methods of decision trees induced from data. We observe that those methods exploiting an independent pruning set do not perform uniformly better than the others. Furthermore, a clear definition of bias towards overpruning and underpruning is exploited in order to interpret empirical data concerning the size of the simplified trees.File in questo prodotto:
Non ci sono file associati a questo prodotto.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.