Data-Driven Electricity Price Risk Assessment for Spot Market

Table 1

Hyperparameter settings of the data-driven methods.

Case	Method	Hyperparameter settings
IEEE 30	M1	3 layers, 100 nodes per layer, and 200 epochs; learning rate = 0.0001, branch size = 500
	M2	3 layers, 100 nodes per layer, and 200 epochs; learning rate = 0.0001, branch size = 500
	M3	500 nodes, 50 reduced hidden nodes, and 2 epochs
	M4	m(x) = 0, C(·, ·) = CSE(·, ·), 100 epochs
	M5	m(x) = 0, C(·, ·) = CSE(·, ·), 100 epochs
IEEE 118	M1	3 layers, 300 nodes per layer, and 300 fine-tuning epochs; learning rate = 0.0001; branch size = 500
	M2	3 layers, 300 nodes per layer, and 300 fine-tuning epochs; learning rate = 0.0001; branch size = 100
	M3	1000 nodes, 100 reduced hidden nodes, and 4 epochs
	M4	m(x) = 0, C(·, ·) = CSE(·, ·), 100 epochs
	M5	m(x) = 0, C(·, ·) = CSE(·, ·), 100 epochs