Using improved support vector regression to predict the transmitted energy consumption data by distributed wireless sensor network

EURASIP Journal on Wireless Communications and Networking

Table 4 Optimized parameters of RF, GBR, and SVR

Methods	Parameters	Parameter description	Terminal values
RF	n_estimators	The number of trees in the forest	10
	max_depth	The depth of a tree	7
GBR	max_depth	The depth of a tree	7
	learning_rate	The learning rate	0.19
	n_estimators	The number of trees	40
SV	C	The penalty parameter	150
	ε	Insensitive loss coefficient	0.08
	γ	γ defines how much influence a single training example has. The larger the γ is, the closer other example must be to be affected.	0.1