A New Fast Training Algorithm for Autoencoder Neural Networks based on Extreme Learning Machine

Vásquez-Coronel, José A.; Mora, Marco; Vilches, Karina; Silva-Pavez, Fabian; Torres-Gonzalez, Italo; Barria-Valdevenito, Pedro

doi:10.1109/ica-acca56767.2022.10006276

Search citation statements

Order By: Relevance

Paper Sections

Select...

Introduction1

Citation Types

Supporting

Mentioning

Contrasting

Year Published

2024

Publication Types

Select...

Other1

Relationship

Self Cite0

Independent1

Authors

Journals

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These issues limit the accuracy of statistical approaches when applied to photovoltaic power prediction [4]. In neural network methods, Long Short-Term Memory (LSTM) [5], as a type of nonlinear recurrent network, has advantages over traditional neural networks in terms of temporal correlation and long-term memory. It has been widely applied in recent years.…”

Section: Introductionmentioning

confidence: 99%

Photovoltaic power prediction method based on CNN-BIGRU-ATTENTION model

Luo,

He,

Chen

et al. 2024

Eighth International Conference on Energy System, Electricity, and Power (ESEP 2023)

View full text Add to dashboard Cite

This study addresses the challenge of accurately predicting photovoltaic (PV) power output across diverse weather conditions. To tackle this issue, we propose a novel combined prediction model that integrates Convolutional Neural Network (CNN), Bidirectional Gated Recurrent Unit (BIGRU), and Attention mechanisms. First, the Copula method is applied to identify the crucial meteorological features that impact photovoltaic output. Subsequently, K-means clustering is employed to group daily photovoltaic power generation scenarios. Following this, the CNN-BIGRU-ATTENTION model is established and applied for power prediction across various clustering scenarios. Finally, we conduct a case study utilizing photovoltaic power generation data from Australia. The results unequivocally demonstrate a significant enhancement in prediction accuracy achieved by the proposed method.

show abstract

Section: Introductionmentioning

confidence: 99%