Search

Electricity price forecasts have become a fundamental factor affecting the decision-making of all market participants. Extreme price volatility has forced market participants to hedge against volume risks and price movements. Hence, getting an accurate price forecast from a few hours to a few days ahead is very important and very challenging due to various factors. This paper proposes an integrated long-term recurrent convolutional network (ILRCN) model to predict electricity prices considering the majority of contributing attributes to the market price as input. The proposed ILRCN model combines the functionalities of a convolutional neural network and long short-term memory (LSTM) algorithm along with the proposed novel conditional error correction term. The combined ILRCN model can identify the linear and nonlinear behavior within the input data. ERCOT wholesale market price data along with load profile, temperature, and other factors for the Houston region have been used to illustrate the proposed model. The performance of the proposed ILRCN electricity price forecasting model is verified using performance/evaluation metrics like mean absolute error and accuracy. Case studies reveal that the proposed ILRCN model shows the highest accuracy and efficiency in electricity price forecasting as compared to the support vector machine (SVM) model, fully connected neural network model, LSTM model, and the traditional LRCN model without the conditional error correction stage. [ABSTRACT FROM AUTHOR]

In recent years, hyperspectral image (HSI) classification has become a hot research direction in remote sensing image processing. Benefiting from the development of deep learning, convolutional neural networks (CNNs) have shown extraordinary achievements in HSI classification. Numerous methods combining CNNs and attention mechanisms (AMs) have been proposed for HSI classification. However, to fully mine the features of HSI, some of the previous methods apply dense connections to enhance the feature transfer between each convolution layer. Although dense connections allow these methods to fully extract features in a few training samples, it decreases the model efficiency and increases the computational cost. Furthermore, to balance model performance against complexity, the AMs in these methods compress a large number of channels or spatial resolutions during the training process, which results in a large amount of useful information being discarded. To tackle these issues, in this article, a novel one-shot dense network with polarized attention, namely, OSDN, was proposed for HSI classification. More precisely, since HSI contains rich spectral and spatial information, the OSDN has two independent branches to extract spectral and spatial features, respectively. Similarly, the polarized AMs contain two components: channel-only AMs and spatial-only AMs. Both polarized AMs can use a specially designed filtering method to reduce the complexity of the model while maintaining high internal resolution in both the channel and spatial dimensions. To verify the effectiveness and lightness of OSDN, extensive experiments were carried out on five benchmark HSI datasets, namely, Pavia University (PU), Kennedy Space Center (KSC), Botswana (BS), Houston 2013 (HS), and Salinas Valley (SV). Experimental results consistently showed that the OSDN can greatly reduce computational cost and parameters while maintaining high accuracy in a few training samples. [ABSTRACT FROM AUTHOR]