Short-Time Wind Speed Forecast Using Artificial Learning-Based Algorithms

Table 2

Acronyms and notations used.

Category	Items/symbols	Description
Acronyms	ANN	Artificial neural network
	CNN	Convolutional neural network
	LSTM	Long short-term memory
	ConvLSTM	Convolutional LSTM hybrid model
	SVM	Support vector machine
	RE	Renewable energy
	RNN	Recurrent neural network
	EWT	Empirical wavelet transformation
	ENN	Elman neural network
	FC-LSTM	Fully connected-long short-term memory
	FCN-LSTM	Long short-term memory fully convolutional network
	TSP	Time-series prediction
	UKF	Unscented Kalman filter
	SVR	Support vector regression
	SVRM	Support vector regression machine
	EO	Extremal optimization
	MAE	Mean absolute error
	RMSE	Root mean square error
	MAPE	Mean absolute percentage error
	R^2	R-squared
	WT	Wavelet transform
	GA	Genetic algorithm
	GAWNN	Genetic algorithm of wavelet neural network
	WNN	Wavelet neural network
	MLOS	Multi-lags-one-step
	VGP	Vanishing gradient problem
	LM	Levenberg–Marquardt
	RBF	Radial basis function
Notations		Forget gate
		The cell state
		Input gate
		Current input data
		The previous hidden output
		Input to cell c
		Memory cell
		Input to cell c
		Input gate
		Past cell status
		Output gate
		Hidden state
		Matrix multiplication
		An elementwise multiplication
		Weight
		The input to the cell
		Nonlinear function
		The jth hidden neuron
		Number of inputs to the network
	m	Number of hidden neurons
		The connection weight from the ith input node to the jth hidden node
		i-step behind previous wind speed
		The activation function in the hidden layer
		The connection weight from the jth hidden node to the output node
		The predicted wind speed at the kth sampling moment
		The activation function for the output layer
		Actual wind speed
		Input vector
		Output vector
		Regularized function
		A function that describes the correlation between inputs and outputs.
		Preknown function
		Structure risk
		The regression coefficient vector
		Bias term
		Punishment coefficient
		The ε-insensitive loss function
	ε	Threshold
		Slack variables that let constraints feasible
		The Lagrange multipliers
		The kernel function
		The weight matrix
		Convolution operation
		Bias vectors
		Hadamard product
		Hidden state
		Current wind speed measure
		Previous wind speed measure
		Future wind speed measure