no code implementations • 14 Mar 2024 • Georgia Papacharalampous, Hristos Tyralis, Nikolaos Doulamis, Anastasios Doulamis
This demonstrates the potential of stacking to improve probabilistic predictions in spatial interpolation and beyond.
no code implementations • 13 Nov 2023 • Georgia Papacharalampous, Hristos Tyralis, Nikolaos Doulamis, Anastasios Doulamis
Compared to QR, LightGBM showed improved performance with respect to the quantile scoring rule by 11. 10%, followed by QRF (7. 96%), GRF (7. 44%), GBM (4. 64%) and QRNN (1. 73%).
no code implementations • 9 Jul 2023 • Georgia Papacharalampous, Hristos Tyralis, Nikolaos Doulamis, Anastasios Doulamis
In this context, satellite precipitation and topography data are the predictor variables, and gauged-measured precipitation data are the dependent variables.
no code implementations • 17 Jun 2023 • Hristos Tyralis, Georgia Papacharalampous, Nilay Dogulu, Kwok P. Chun
The main idea is to train a deep learning algorithm with the Huber quantile regression function, which is consistent for the Huber quantile functional.
no code implementations • 2 Feb 2023 • Hristos Tyralis, Georgia Papacharalampous, Nikolaos Doulamis, Anastasios Doulamis
To improve precipitation estimates, machine learning is applied to merge rain gauge-based measurements and gridded satellite precipitation products.
no code implementations • 31 Dec 2022 • Georgia Papacharalampous, Hristos Tyralis, Anastasios Doulamis, Nikolaos Doulamis
Still, information on which tree-based ensemble algorithm to select for correcting satellite precipitation products for the contiguous United States (US) at the daily time scale is missing from the literature.
no code implementations • 17 Dec 2022 • Georgia Papacharalampous, Hristos Tyralis, Anastasios Doulamis, Nikolaos Doulamis
To provide results that are generalizable and to contribute to the delivery of best practices, we here compare eight state-of-the-art machine learning algorithms in correcting satellite precipitation data for the entire contiguous United States and for a 15-year period.
no code implementations • 17 Sep 2022 • Hristos Tyralis, Georgia Papacharalampous
Predictions and forecasts of machine learning models should take the form of probability distributions, aiming to increase the quantity of information communicated to end users.
no code implementations • 17 Jun 2022 • Georgia Papacharalampous, Hristos Tyralis
Several machine learning concepts and methods are notably relevant towards addressing the major challenges of formalizing and optimizing probabilistic forecasting implementations, as well as the equally important challenge of identifying the most useful ones among these implementations.
no code implementations • 13 Apr 2022 • Georgia Papacharalampous, Hristos Tyralis, Yannis Markonis, Petr Maca, Martin Hanel
Detailed feature investigations and comparisons across climates, continents and time series types can progress our understanding and modelling ability of the Earth's hydroclimate and its dynamics.
no code implementations • 13 Apr 2022 • Georgia Papacharalampous, Hristos Tyralis
Precipitation and temperature features (e. g., the spectral entropy, seasonality strength and lag-1 autocorrelation of the precipitation time series, and the stability and trend strength of the temperature time series) were found to be useful predictors of many streamflow features.
no code implementations • 2 Dec 2021 • Georgia Papacharalampous, Hristos Tyralis, Yannis Markonis, Martin Hanel
A comprehensive understanding of the behaviours of the various geophysical processes and an effective evaluation of time series (else referred to as "stochastic") simulation models require, among others, detailed investigations across temporal scales.
no code implementations • 25 Jul 2021 • Georgia Papacharalampous, Hristos Tyralis, Ilias G. Pechlivanidis, Salvatore Grimaldi, Elena Volpi
Statistical analyses and descriptive characterizations are sometimes assumed to be offering information on time series forecastability.
no code implementations • 16 Apr 2021 • Georgia Papacharalampous, Andreas Langousis
The results mostly favour the practical systems designed using the linear boosting algorithm, probably due to the presence of trends in the urban water flow time series.
no code implementations • 1 Apr 2020 • Hristos Tyralis, Georgia Papacharalampous
Machine learning algorithms have been extensively exploited in energy research, due to their flexibility, automation and ability to handle big data.
no code implementations • 2 Jan 2020 • Georgia Papacharalampous, Hristos Tyralis
In this work, we present and appraise a new simple and flexible methodology for hydrological time series forecasting.
no code implementations • 9 Sep 2019 • Hristos Tyralis, Georgia Papacharalampous, Andreas Langousis
Based on the obtained large-scale results, we propose super learning for daily streamflow forecasting.