When fitting statistical models to variables in geoscientific disciplines such as hydrology, it is a customary practice to regionalize - to divide a large spatial domain into multiple regions and study each region separately - instead of fitting a single model on the entire data (also known as unification). Traditional wisdom in these fields suggests that models built for each region separately will have higher performance because of homogeneity within each region. However, by partitioning the training data, each model has access to fewer data points and cannot learn from commonalities between regions. Here, through two hydrologic examples (soil moisture and streamflow), we argue that unification can often significantly outperform regionalization in the era of big data and deep learning (DL). Common DL architectures, even without bespoke customization, can automatically build models that benefit from regional commonality while accurately learning region-specific differences. We highlight an effect we call data synergy, where the results of the DL models improved when data were pooled together from characteristically different regions. In fact, the performance of the DL models benefited from more diverse rather than more homogeneous training data. We hypothesize that DL models automatically adjust their internal representations to identify commonalities while also providing sufficient discriminatory information to the model. The results here advocate for pooling together larger datasets, and suggest the academic community should place greater emphasis on data sharing and compilation.
Leveraging Exogenous Signals for Hydrology Time Series Forecasting
College of Computing(计算学院);Georgia Institute of Technology(佐治亚理工学院);University of Virginia(弗吉尼亚大学);School of Data Science(数据科学学院);Computer Science(计算机科学);Biocomplexity Institute(生物复杂性研究所)
Pix2Streams: Dynamic Hydrology Maps from Satellite-LiDAR Fusion
Frontier Development Lab(前沿发展实验室);University of Valencia(瓦伦西亚大学);Arizona State University(亚利桑那州立大学);NASA Ames Research Center(美国国家航空航天局艾姆斯研究中心);USGS(美国地质调查局);RSS-Hydro;University of Bristol(布里斯托尔大学);University of Oxford(牛津大学)
Deep Learning Foundation and Pattern Models: Challenges in Hydrological Time Series
Computer Science(计算机科学);University of Virginia(弗吉尼亚大学);College of Computing(计算学院);Georgia Institute of Technology(佐治亚理工学院);Biocomplexity Institute(生物复杂性研究所)
Massive feature extraction for explaining and foretelling hydroclimatic time series forecastability at the global scale
Roma Tre University(罗马第三大学);National Technical University of Athens(雅典国立技术大学);Hellenic Air Force(希腊空军);Swedish Meteorological and Hydrological Institute(瑞典气象水文研究所);University of Tuscia(图西亚大学);New York University(纽约大学)
Rank-based change-point analysis for long-range dependent time series
我们考虑基于秩统计量的变点检验来检验长期相关观测值的结构变化。在平稳时间序列的假设下,在变点高度随变点高度减小而变化的假设下,导出了相应的检验统计量的渐近分布。为此,证明了序列经验过程在具有加权上确界范数的两参数Skorohod空间中的一致约化原理。此外,我们比较了秩检验的效率所产生的考虑不同的评分函数。在高斯性下,基于秩的检验相对于CuSum检验的渐近相对效率为1,与评分函数无关。在秩基变点检验的实际应用中,我们建议将自归一化秩统计量与二次抽样相结合。理论结果是伴随着模拟研究,特别是,允许比较秩检验从不同的评分函数。对于基于秩的变点检验的有限样本性能,Van der Waerden秩检验被证明在广泛的情况下是有利的。最后,我们分析了经济,水文和网络流量监测的数据集的结构变化,并将我们的结果与以前的数据分析进行比较。
On The Estimation of the Hurst Exponent Using Adjusted Rescaled Range Analysis, Detrended Fluctuation Analysis and Variance Time Plot: A Case of Exponential Distribution
iTFKAN: Interpretable Time Series Forecasting with Kolmogorov-Arnold Network
Hong Kong Polytechnic University(香港理工大学);Northwestern Polytechnical University(西北工业大学);Sun Yat-sen University(中山大学);Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州))
GateTS: Versatile and Efficient Forecasting via Attention-Inspired routed Mixture-of-Experts
Department of Artificial Intelligence, Lviv Polytechnic National University(人工智能系,利沃夫工业大学);Department of Electrical Engineering, Eindhoven University of Technology(电气工程系,埃因霍温理工大学);The Bartlett School of Sustainable Construction, University College London(可持续建设学院,伦敦大学学院)
pyFAST: A Modular PyTorch Framework for Time Series Modeling with Multi-source and Sparse Data
College of Computer Engineering, Jimei University(嘉应大学计算机工程学院);Chengyi College, Jimei University(嘉应大学 Chengyi 学院);Department of Technology, Management and Economics, Technical University of Denmark(丹麦技术大学技术、管理与经济系)
Fusing Optical and SAR time series for LAI gap filling with multioutput Gaussian processes
Universitat de València(瓦伦西亚大学)
卫星光学信息的可用性往往受到云的自然存在的阻碍,这对许多应用来说可能是个问题。农田上空持续的云层会掩盖作物生长的关键阶段,导致产量预测不可靠。合成孔径雷达(SAR)提供全天候图像,有可能克服这一限制,但鉴于其对不同表面特性的高度和独特的敏感性,SAR和光学数据的融合仍然是一个公开的挑战。在这项工作中,我们提出了使用多输出高斯过程(MOGP)回归,机器学习技术,自动学习多传感器时间序列之间的统计关系,检测植被覆盖区,SAR光学图像之间的协同作用是有利可图的。为此,我们使用哨兵-1雷达植被指数(RVI)和哨兵-2叶面积指数(LAI)的时间序列在西北伊比利亚半岛的研究区域。通过对MOGP训练模型的物理解释,我们展示了它使用与RVI共享的信息提供LAI估计的能力,即使在多云期间,这保证了解决方案始终与实际测量结果相关联。结果表明MOGP的优势,特别是对于长的数据间隙,其中基于光学的方法是出了名的失败。应用于整个植被覆盖的留一影像评估技术表明,MOGP预测在短时间间隔内(R $^2$为74% vs 68%,RMSE为0.4 vs 0.44 $[m^2m^{-2}]$),特别是在长时间间隔内(R $^2$为33% vs 12%,RMSE为0.5 vs 1.09 $[m^2m^{-2}]$),提高了标准GP估计。
FAIM: Frequency-Aware Interactive Mamba for Time Series Classification
School of Artificial Intelligence, OPtics and ElectroNics (iOPEN), Northwestern Polytechnical University(人工智能学院、光学与电子学(iOPEN)、西北工业大学);Institute of Artificial Intelligence (TeleAI), China Telecom(人工智能研究院(TeleAI)、中国电信)
Time Distributed Deep Learning Models for Purely Exogenous Forecasting: Application to Water Table Depth Prediction using Weather Image Time Series
Computer Science Department - University of Turin(都灵大学计算机科学系);LISTIC Laboratory - Université Savoie Mont Blanc(萨瓦伊-蒙布莱大学LISTIC实验室);Interuniversity Department of Regional and Urban Studies and Planning, Politecnico di Torino and University of Turin(都灵理工学院和都灵大学区域与城市研究与规划联合学院)
Deep Learning for Wildfire Risk Prediction: Integrating Remote Sensing and Environmental Data
Schulich School of Engineering, Department of Geomatics Engineering, University of Calgary(Schulich 工程学院,测绘工程系,卡尔加里大学);Department of Geography and Environmental Management, University of Waterloo(地理与环境管理系,滑铁卢大学);Department of Systems Design Engineering, University of Waterloo(系统设计工程系,滑铁卢大学);School of Emergency Management, Nanjing University of Information Science and Technology(应急管理学院,南京信息工程大学);Division of Geoinformatics, KTH Royal Institute of Technology(地理信息学系,皇家理工学院);School of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology(遥感与测绘工程学院,南京信息工程大学);Department of Software Engineering, University of Waterloo(软件工程系,滑铁卢大学);Department of Management Science and Engineering, University of Waterloo(管理科学与工程系,滑铁卢大学)