Cited by Spatio-temporal long short-term memory neural network with seasonal-trend decomposition for ambient air pollutant forecasting

Recurrent Fourier-Kolmogorov Arnold Networks for photovoltaic power forecasting DOI

Desheng Rong, Zhongbao Lin,

Guomin Xie

и другие.

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Фев. 8, 2025

Accurate day-ahead forecasting of photovoltaic (PV) power generation is crucial for system scheduling. To overcome the inaccuracies and inefficiencies current PV models, this paper introduces Recurrent Fourier-Kolmogorov Arnold Network (RFKAN). Initially, recurrent kernel nodes are employed to investigate interdependencies within sequences. Subsequently, Fourier series applied extract periodic features, enhancing accuracy training speed. Ablation studies conducted using data from a plant in Tieling City, Liaoning Province, validate effectiveness these two structural enhancements. Comparative experiments with baseline state-of-the-art models further underscore efficiency RFKAN. The results indicate that RFKAN achieves best performance grid depth 100 an input sequence length 2, reducing RMSE MAE by at least 5%, increasing CORR 2%, decreasing time 24% compared advanced models.

Язык: Английский

Процитировано

Domain perceptive-pruning and fine-tuning the pre-trained model for heterogeneous transfer learning in cross domain prediction DOI

Dan Yang, Xin Peng, Xiaolong Wu

и другие.

Expert Systems with Applications, Год журнала: 2024, Номер 260, С. 125215 - 125215

Опубликована: Сен. 19, 2024

Язык: Английский

Процитировано

Time Series Forecasting Method Based on Multi-Scale Feature Fusion and Autoformer DOI

Xiangkai Ma, Huaxiong Zhang

Applied Sciences, Год журнала: 2025, Номер 15(7), С. 3768 - 3768

Опубликована: Март 29, 2025

Accurate time series forecasting is crucial in fields such as business, finance, and meteorology. To achieve more precise predictions effectively capture the potential cycles stochastic characteristics at different scales series, this paper optimizes network structure of Autoformer model. Based on multi-scale convolutional operations, a feature fusion proposed, combined with date–time encoding to build MD–Autoformer model, which enhances model’s ability information scales. In tasks across four fields—apparel sales, meteorology, disease—the proposed method achieved lowest RMSE MAE. Additionally, ablation experiments demonstrated effectiveness reliability method. Combined TPE Bayesian optimization algorithm, prediction error was further reduced, providing reference for future research methods.

Язык: Английский

Процитировано

Enhancing PM2.5 Forecasting Using Video-Based Spatiotemporal Models and Cyclical Encoding DOI

Jitti Pranonsatit,

Kritchart Wongwailikhit,

Pisut Painmanakul

и другие.

Communications in computer and information science, Год журнала: 2025, Номер unknown, С. 357 - 372

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

Fusing Physics into Machine Learning for Complex Air Pollutant DOI

Yueyang Sun,

Chase Wu,

Nan Guo

и другие.

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

Air Quality Index Prediction through TimeGAN Data Recovery and PSO-Optimized VMD-Deep Learning Framework DOI

Kenan Wang,

Tianning Yang,

Shanshan Kong

и другие.

Applied Soft Computing, Год журнала: 2024, Номер unknown, С. 112626 - 112626

Опубликована: Дек. 1, 2024

Язык: Английский

Процитировано

Spatio-temporal long short-term memory neural network with seasonal-trend decomposition for ambient air pollutant forecasting DOI

Rui Zhang, Norhashidah Awang, Feng Jing

и другие.

Earth Science Informatics, Год журнала: 2024, Номер 18(1)

Опубликована: Дек. 26, 2024

Язык: Английский

Процитировано