Long term prediction of inflow using the supervised learning (CROSBI ID 246513)
Prilog u časopisu | izvorni znanstveni rad | međunarodna recenzija
Podaci o odgovornosti
Berbić, Jadran ; Ocvirk, Eva
engleski
Long term prediction of inflow using the supervised learning
For the purposes of long-term (on time basis of one month) planning and management of water resources systems, long term prediction of inflow is needed. In the last two decades, usage of machine learning is becoming popular in the field of water resources systems management, whether for real time, short term, mid term or long term predictions of hydrological variables. Especially interesting is the usage of supervised learning, defined as the type of machine learning used for model development based on given data, which enables prediction and extrapolation on so far unseen examples. Supervised learning models are able to use arbitrarily huge amount of variables for model development and forecasting. Mentioned facts make the problematics interesting from the climate change point of view, as also from the view of model development for assistance in water resources systems planning and management. By using the models developed on historical data it is able to predict inflows in conditions of future scenario from climate models and get insight in future hydrological conditions and systems management efficiency. Besides classicaly used rainfall for runoff modelling, other meteorological variables, if are on disposition, could be included. As searching for appropriate way of prediction and supervised learning models architecture is not an easy task, necessary step is reviewing the literature about long-term predictions. Therefore, the insight in previous research of long term prediction of inflow using the supervised learning is given in the paper. In the literature review are also included researches with considerations of climate change influence on water resources systems, based on the predictions using the supervised learning.
supervised learning, long-term prediction, water resources systems, climate change
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano