Wind Power Probability Density Prediction Based on Quantile Regression Model of Dilated Causal Convolutional Neural Network

Abstract

Aiming at the wind power prediction problem, a wind power probability prediction method based on the quantile regression of a dilated causal convolutional neural network is proposed. With the developed model, the Adam stochastic gradient descent technique is utilized to solve the cavity parameters of the causal convolutional neural network under different quantile conditions and obtain the probability density distribution of wind power at various times within the following 200 hours. The presented method can obtain more useful information than conventional point and interval predictions. Moreover, a prediction of the future complete probability distribution of wind power can be realized. According to the actual data forecast of wind power in the PJM network in the United States, the proposed probability density prediction approach can not only obtain more accurate point prediction results, it also obtains the complete probability density curve prediction results for wind power. Compared with two other quantile regression methods, the developed technique can achieve a higher accuracy and smaller prediction interval range under the same confidence level.

Measurement of Systemic Risk Based on the QRDCCNN Model

“…It uses dilated causal convolution for padding, reducing the number of parameters and expanding the model's receptive field while ensuring the causal consistency of financial time series. [18].…”

“…The dimension rules of the output shape of the input layer and convolution layer in Table 3 can be referred to in Table 4. Additionally, the number of parameters for each convolution layer in Table 3 can be calculated using formula (18), which is mathematically expressed as follows:…”

“…Building on the QRCNN model, Yang et al [18] introduced the dilated causal convolution from the WaveNet model into the QRCNN model and proposed the Quantile Regression Dilated Causal Convolution Neural Network (QRDCCNN) model. Dilated causal convolution, originating in speech recognition, has shown superior performance in audio synthesis by ensuring temporal causality and expanding the receptive field [19].…”