相关参考文献
注意:仅列出部分参考文献,下载原文获取全部文献信息。
Article
Operations Research & Management Science
Ayse Ozmen
Summary: This paper examines the application of the multivariate adaptive regression splines (MARS) model and the least absolute shrinkage and selection operator (LASSO) model in forecasting natural gas demand. The results show that MARS outperforms LASSO and multiple-linear regression (LR) in short-term forecasting, but the performance of all three models deteriorates in long-term forecasting.
ANNALS OF OPERATIONS RESEARCH
(2023)
Article
Energy & Fuels
Jia Ding et al.
Summary: Natural gas is essential for power generation, urban heating, and manufacturing. Accurate forecasting of its consumption is crucial to ensure a reliable supply. This paper proposes a novel method, Dual Convolution with Seasonal Decomposition Network, which outperforms existing methods in terms of prediction accuracy and sensitivity regardless of different time intervals. The method can be applied to residential quarters, cities, or even countries in practical circumstances.
Article
Automation & Control Systems
Xin Ma et al.
Summary: A novel wavelet kernel-based grey system model is proposed to address the challenges of collecting urban natural gas consumption data in China, high nonlinearity and complex features of the data sets, and the difficulty of accurate forecasting for mid-small cities with small samples. A complete computational algorithm is presented, and the optimal hyperparameters are selected through a hold-out cross validation-based grid-search scheme. Three case studies based on real-world data sets of urban natural gas consumption in Kunming China demonstrate the model's superior performance compared to 15 other time series forecasting models, highlighting its importance and potential in similar applications.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Thermodynamics
Jolanta Szoplik et al.
Summary: This paper presents the results of natural gas composition forecasting using the MLP model of artificial neural network. The model includes calendar and weather factors that indirectly affect the gas composition. The best quality MLP 18-65-5 network was selected based on correlation coefficient and MAPE forecast errors. Forecasts were made for the next calendar year with an average MAPE forecast error of 3.356%.
Article
Environmental Studies
Manzhi Liu et al.
Summary: This study combines structural decomposition analysis and input-output subsystem analysis to study the key drivers of China's carbon dioxide emissions. By analyzing the influence of various factors on carbon dioxide emissions, it is found that intersectoral pulling effects, particularly from the Construction sector, contribute to the increase in emissions. The study also highlights the importance of changing the current energy structure to reduce emissions, with the technical progress and energy structure optimization scenarios predicted to achieve peak emissions in 2025 and 2030, respectively.
ENVIRONMENTAL IMPACT ASSESSMENT REVIEW
(2023)
Article
Computer Science, Artificial Intelligence
Kai Zhang et al.
Summary: In grey system theory, the performance of a grey forecasting model depends on efficiently measuring grey information from data. This study proposed a probabilistic accumulation operator-based grey forecasting model (PGM(1,1)) that uses a Bernoulli distribution to simulate valid/invalid grey information. PGM(1,1) achieved state-of-the-art performance by not being affected by invalid grey information, as demonstrated through comprehensive analysis and comparison with other models on five public datasets.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Environmental Sciences
Mengyao Xia et al.
Summary: Corporate carbon performance is crucial for achieving corporate sustainability. Understanding the factors influencing carbon emissions is essential in promoting carbon performance. Using the CDP database, we use the LASSO regression model and fixed effects model to identify and rank the determinants of carbon emissions. Our findings show that Capx is included in all carbon contexts, while financial-level factors play a greater role in Scope 1 and Scope 2 emissions. For Scope 3, internal incentive policies and emission reduction behaviors are important, and the debt-paying ability of financial-level factors serves as a vital indicator for relative carbon emissions.
ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH
(2023)
Article
Chemistry, Physical
Zhaoqiu Lyu et al.
Summary: This paper discusses the feasibility of using a novel machine learning approach with K-fold cross-validation to predict the torsional strength of Reinforced Concrete (RC) beams. The study discovers that by optimizing neural network parameters and utilizing K-fold cross-validation and genetic algorithms, the accuracy of the prediction model can be improved.
Article
Energy & Fuels
Mohamad Hossein Safiyari et al.
Summary: This study examines forecasting models for natural gas demand and finds that the multi-layer perceptron model performs the best in predicting natural gas demand in the residential sector of Tehran province.
Article
Green & Sustainable Science & Technology
Q. R. Cao et al.
Summary: China's carbon-reduction policies have negatively impacted provincial industrial competitiveness, with a significant effect in the eastern regions and a smaller effect in the central regions. The carbon emission trading policy has generally had a positive impact on industrial competitiveness. Different types of industries have been affected differently, with extractive, low-energy-consuming, and high-tech manufacturing industries experiencing harm to competitiveness. However, there has been no discernible impact on high-energy-consuming industries. The carbon trading policy has improved the competitiveness of high energy-consuming industries provincially.
Article
Economics
Sheng Zeng et al.
Summary: China has made significant progress in optimizing its energy structure, and is expected to achieve its carbon emission targets in the future and complete the carbon emission target per unit of GDP ahead of schedule.
Proceedings Paper
Green & Sustainable Science & Technology
Jingjing Min et al.
Summary: The study found that the EMD method is more suitable for decomposing the social and meteorological demand of natural gas, and there is a close relationship between natural gas demand in the heating season and meteorological factors. The newly proposed EMD_BP prediction model accurately fits the changing trend of natural gas time series.
2020 ASIA CONFERENCE ON GEOLOGICAL RESEARCH AND ENVIRONMENTAL TECHNOLOGY
(2021)
Article
Thermodynamics
Weijie Zhou et al.
Article
Energy & Fuels
Tomas Balezentis et al.
Article
Environmental Sciences
Lihao Gao et al.
Article
Thermodynamics
Faheemullah Shaikh et al.
Article
Engineering, Electrical & Electronic
Faheemullah Shaikh et al.
INTERNATIONAL JOURNAL OF ELECTRICAL POWER & ENERGY SYSTEMS
(2016)
Article
Thermodynamics
L. Zhu et al.
Article
Energy & Fuels
Vincenzo Bianco et al.