It is important to consider the changing states in hedging.The Markov regime-switching dynamic correlation multivariate stochastic volatility( MRS-DC-MSV) model was proposed to solve this issue. DC-MSV model and MRS-D...It is important to consider the changing states in hedging.The Markov regime-switching dynamic correlation multivariate stochastic volatility( MRS-DC-MSV) model was proposed to solve this issue. DC-MSV model and MRS-DC-MSV model were used to calculate the time-varying hedging ratios and compare the hedging performance. The Markov chain Monte Carlo( MCMC) method was used to estimate the parameters. The results showed that,there were obviously two economic states in Chinese financial market. Two models all did well in hedging,but the performance of MRS-DCMSV model was better. It could reduce risk by nearly 90%. Thus,in the hedging period,changing states is a factor that cannot be neglected.展开更多
Short-term offshore wind power forecasting is crucial for stable power system operations.However,accurate forecasting is hindered by multivariate interactions that generate multi-scale repetitive patterns and obscure ...Short-term offshore wind power forecasting is crucial for stable power system operations.However,accurate forecasting is hindered by multivariate interactions that generate multi-scale repetitive patterns and obscure cross-variate correlations.In this paper,we propose a hierarchical framework,the Time-Frequency Clustering Inverted Transformer,for multivariate offshore wind power forecasting.First,a Time-Frequency Clustering component applies Wavelet Packet Decomposition to each series and quantifies sub-series similarity by overall activity and evolutionary trend,grouping repetitive patterns into structured clusters.Second,an inverted Transformer captures multivariate correlations within clusters by treating time points of individual sub-series as variate tokens,enabling self-attention to focus on multivariate correlations rather than temporal dependencies.On two real-world offshore wind datasets(horizons 8-48 h),our proposed framework reduces MSE/MAE by 14.11%and outperforms 12 recognised baselines(e.g.,PatchTST,TimesNet),with the advantage persisting even when the TFC component is applied to the baselines.Moreover,our method demonstrates remarkable generalisability on three public datasets(Solar-Energy,Traffic,and ECL),reducing MSE/MAE by 7.36%.These results indicate that associating repetitive patterns with attention to cross-variate structure materially improves multivariate offshore wind power forecasting.展开更多
This work developed a statistical correlation between groundwater’s high iron content in the four hydrogeological domains of the State of Bahia,Brazil,and the environmental attributes of climate,lithology,soil,and ve...This work developed a statistical correlation between groundwater’s high iron content in the four hydrogeological domains of the State of Bahia,Brazil,and the environmental attributes of climate,lithology,soil,and vegetation.From 3539 wells,flow test≥1 m3∙h−1,drilling period 2003-2013,940 wells with high iron content(>0.3 mg/L)were used in this study.All groundwater samples came from new wells soon after the drilling,well construction,and a long pumping time for their development:24 hours for sedimentary aquifers and 12 hours for karstic,crystalline,and metasedimentary aquifers.The applica-tion of Pearson and Spearman linear regression to seventeen physicochemical parameters(SPSS V.12)resulted in no correlations between iron and fourteen parameters,indicating no common origin between those parameters and iron.Only color and turbidity presented correlations>0.20 with iron.After spati alizing the 940 values of iron concentration(ArcGIS V.9)on the maps of each environmental attribute,grades 1-5 were given to the variables of each attrib-ute based on the largest iron concentration value.The grades allowed the ap-plication of multivariable methods PCA and FA(SPSS V.12).The PCA indi cated two factors explaining 59.52%of the total variance,closely attending the recommended minimum of 60%.The significant factor weights from the ap-plication of FA were:in Factor 1,soil,−0.71;vegetation,−0.68;and lithology,−0.52;and in Factor 2,climate,+0.74.Indeed,in the crystalline and metased-imentary domains with mafic-ultramafic rocks rich in iron,percentages of wells,53.3%-66.7%,occurred in iron-rich soils;of 49.8%-59.8%in humid to dry forest and of 55.3%-86.8%in humid to sub-humid climate.While,for the sedimentary domain(primarily sandstones)and karstic domain(carbonate rocks)poor in iron content percentages of wells,80.9%-100%occurred in iron-rich soils,57.0%-61.8%in humid to dry forest,and 58.6%-62.4%in sub-humid to dry and semi-arid climate.These results indicated that,although lithology is a determinant for high dissolved iron content in the state of Bahia groundwater,this attribute alone(factor weight−0.52)cannot explain the whole phenomenon.The present work,using multivariable analysis with geo-spatial mapping of high iron content on top of environmental attributes,re-vealed the role of each environmental attribute in groundwater’s high iron con-tent.For the governmental drilling well company and its groundwater manag-ers,this knowledge will result in better well locations and a reduction of both well and economic losses,as the long-term maintenance cost for the treatment process due to high iron content is prohibitive for rural municipalities.展开更多
Correlated multivariate time series prediction is an effective tool for discovering the chang rules of temporal data,but it is challenging tofind these rules.Recently,deep learning methods have made it possible to pred...Correlated multivariate time series prediction is an effective tool for discovering the chang rules of temporal data,but it is challenging tofind these rules.Recently,deep learning methods have made it possible to predict high-dimensional and complex multivariate time series data.However,these methods cannot capture or predict potential mutation signals of time series,leading to a lag in data prediction trends and large errors.Moreover,it is difficult to capture dependencies of the data,especially when the data is sparse and the time intervals are large.In this paper,we proposed a prediction approach that leverages both propagation dynamics and deep learning,called Rolling Iterative Prediction(RIP).In RIP method,the Time-Delay Moving Average(TDMA)is used to carry out maximum likelihood reduction on the raw data,and the propagation dynamics model is applied to obtain the potential propagation parameters data,and dynamic properties of the correlated multivariate time series are clearly established.Long Short-Term Memory(LSTM)is applied to capture the time dependencies of data,and the medium and long-term Rolling Iterative Prediction method is established by alternately estimating parameters and predicting time series.Experiments are performed on the data of the Corona Virus Disease 2019(COVID-19)in China,France,and South Korea.Experimental results show that the real distribution of the epidemic data is well restored,the prediction accuracy is better than baseline methods.展开更多
基金National Natural Science Foundation of China(No.71401144)
文摘It is important to consider the changing states in hedging.The Markov regime-switching dynamic correlation multivariate stochastic volatility( MRS-DC-MSV) model was proposed to solve this issue. DC-MSV model and MRS-DC-MSV model were used to calculate the time-varying hedging ratios and compare the hedging performance. The Markov chain Monte Carlo( MCMC) method was used to estimate the parameters. The results showed that,there were obviously two economic states in Chinese financial market. Two models all did well in hedging,but the performance of MRS-DCMSV model was better. It could reduce risk by nearly 90%. Thus,in the hedging period,changing states is a factor that cannot be neglected.
基金supported in part by the Department of Human Resources and Social Security of Guangxi Zhuang Autonomous Region(grant number 202401950)in part by the Department of Science and Technology of Guangxi Zhuang Autonomous Region(grant number 2024JJB170087)+2 种基金in part by the National Natural Science Foundation of China(grant number 62371144)in part by the Guangxi Science and Technology Major Program(grant number AA23073019)in part by the The Project of the Philosophy and Social Sciences Fund of Jiangsu Province(grant number 24GLC023).
文摘Short-term offshore wind power forecasting is crucial for stable power system operations.However,accurate forecasting is hindered by multivariate interactions that generate multi-scale repetitive patterns and obscure cross-variate correlations.In this paper,we propose a hierarchical framework,the Time-Frequency Clustering Inverted Transformer,for multivariate offshore wind power forecasting.First,a Time-Frequency Clustering component applies Wavelet Packet Decomposition to each series and quantifies sub-series similarity by overall activity and evolutionary trend,grouping repetitive patterns into structured clusters.Second,an inverted Transformer captures multivariate correlations within clusters by treating time points of individual sub-series as variate tokens,enabling self-attention to focus on multivariate correlations rather than temporal dependencies.On two real-world offshore wind datasets(horizons 8-48 h),our proposed framework reduces MSE/MAE by 14.11%and outperforms 12 recognised baselines(e.g.,PatchTST,TimesNet),with the advantage persisting even when the TFC component is applied to the baselines.Moreover,our method demonstrates remarkable generalisability on three public datasets(Solar-Energy,Traffic,and ECL),reducing MSE/MAE by 7.36%.These results indicate that associating repetitive patterns with attention to cross-variate structure materially improves multivariate offshore wind power forecasting.
文摘This work developed a statistical correlation between groundwater’s high iron content in the four hydrogeological domains of the State of Bahia,Brazil,and the environmental attributes of climate,lithology,soil,and vegetation.From 3539 wells,flow test≥1 m3∙h−1,drilling period 2003-2013,940 wells with high iron content(>0.3 mg/L)were used in this study.All groundwater samples came from new wells soon after the drilling,well construction,and a long pumping time for their development:24 hours for sedimentary aquifers and 12 hours for karstic,crystalline,and metasedimentary aquifers.The applica-tion of Pearson and Spearman linear regression to seventeen physicochemical parameters(SPSS V.12)resulted in no correlations between iron and fourteen parameters,indicating no common origin between those parameters and iron.Only color and turbidity presented correlations>0.20 with iron.After spati alizing the 940 values of iron concentration(ArcGIS V.9)on the maps of each environmental attribute,grades 1-5 were given to the variables of each attrib-ute based on the largest iron concentration value.The grades allowed the ap-plication of multivariable methods PCA and FA(SPSS V.12).The PCA indi cated two factors explaining 59.52%of the total variance,closely attending the recommended minimum of 60%.The significant factor weights from the ap-plication of FA were:in Factor 1,soil,−0.71;vegetation,−0.68;and lithology,−0.52;and in Factor 2,climate,+0.74.Indeed,in the crystalline and metased-imentary domains with mafic-ultramafic rocks rich in iron,percentages of wells,53.3%-66.7%,occurred in iron-rich soils;of 49.8%-59.8%in humid to dry forest and of 55.3%-86.8%in humid to sub-humid climate.While,for the sedimentary domain(primarily sandstones)and karstic domain(carbonate rocks)poor in iron content percentages of wells,80.9%-100%occurred in iron-rich soils,57.0%-61.8%in humid to dry forest,and 58.6%-62.4%in sub-humid to dry and semi-arid climate.These results indicated that,although lithology is a determinant for high dissolved iron content in the state of Bahia groundwater,this attribute alone(factor weight−0.52)cannot explain the whole phenomenon.The present work,using multivariable analysis with geo-spatial mapping of high iron content on top of environmental attributes,re-vealed the role of each environmental attribute in groundwater’s high iron con-tent.For the governmental drilling well company and its groundwater manag-ers,this knowledge will result in better well locations and a reduction of both well and economic losses,as the long-term maintenance cost for the treatment process due to high iron content is prohibitive for rural municipalities.
基金This work was supported by the National Key R&D Program of China under Grant No.2020YFB1710200.
文摘Correlated multivariate time series prediction is an effective tool for discovering the chang rules of temporal data,but it is challenging tofind these rules.Recently,deep learning methods have made it possible to predict high-dimensional and complex multivariate time series data.However,these methods cannot capture or predict potential mutation signals of time series,leading to a lag in data prediction trends and large errors.Moreover,it is difficult to capture dependencies of the data,especially when the data is sparse and the time intervals are large.In this paper,we proposed a prediction approach that leverages both propagation dynamics and deep learning,called Rolling Iterative Prediction(RIP).In RIP method,the Time-Delay Moving Average(TDMA)is used to carry out maximum likelihood reduction on the raw data,and the propagation dynamics model is applied to obtain the potential propagation parameters data,and dynamic properties of the correlated multivariate time series are clearly established.Long Short-Term Memory(LSTM)is applied to capture the time dependencies of data,and the medium and long-term Rolling Iterative Prediction method is established by alternately estimating parameters and predicting time series.Experiments are performed on the data of the Corona Virus Disease 2019(COVID-19)in China,France,and South Korea.Experimental results show that the real distribution of the epidemic data is well restored,the prediction accuracy is better than baseline methods.
