20.500.12556/RUL-173268 Napovedovanje rečnega pretoka z metodami globokega učenja River discharge prediction using deep learning methods Zanesljivo kratko- do srednjeročno napovedovanje pretokov rek je aktualen problem, kar potrjujejo vse pogostejši poplavni dogodki. Napovedovanje pretokov je zahtevno zaradi kompleksnih časovnih in prostorskih dinamik med rečnimi pretoki in meteorološkimi spremenljivkami, kot so padavine. Agencija Republike Slovenije za okolje (ARSO) trenutno uporablja Hidrološki prognosti čni sistem (HPS), ki problem rešuje s tradicionalnimi metodami, ki temeljijo na fizikalnih enačbah. V nalogi predlagamo alternativnen pristop, imenovan SIREN (Synoptic Information-based River Discharge Estimation Network), ki temelji na moderni arhitekturi globokih nevronskih mrež, ki so znane po svojem učinkovitem modeliranju kompleksnih vzorcev. Na realni podatkovni zbirki je bil model SIREN podrobno analiziran in primerjan s HPS ter z modelom HIDRA3river, ki temelji na arhitekturi HIDRA3, globoki nevronski mreži za napovedovanje višin gladine morja. Rezultati so pokazali, da je model SIREN učinkovitejši kot referenčni metodi. Na izbranih metrikah dosega do 62.8% višjo uspešnost kot HPS ter do 30.7% višjo uspešnost kot HIDRA3river, s čimer predstavlja novo stanje tehnike na področju napovedovanja pretokov slovenskih rek. Reliable short- to medium-term river discharge forecasting is a relevant problem, as confirmed by increasingly frequent flood events. Discharge forecasting is challenging due to the complex temporal and spatial dynamics between river discharges and meteorological variables such as precipitation. The Slovenian Environment Agency (ARSO) currently uses the Hydrological Forecasting System (HPS), which addresses the problem with traditional methods based on physical equations. In this work, we propose an alternative approach called SIREN (Synoptic Information-based River Discharge Estimation Network), that is based on a modern architecture of deep neural networks, which are known for their effectiveness in modeling complex patterns. On a real-world dataset, the SIREN model was thoroughly analyzed and compared with HPS and with the HIDRA3river model, which is based on the HIDRA3 architecture, a deep neural network for sea level forecasting. The results showed that the SIREN model is more effective than the reference methods. On selected metrics, it achieves up to 62.8% better performance than HPS and up to 30.7% better performance than HIDRA3river, thus representing the new state of the art in river discharge forecasting for Slovenian rivers. rečni pretok padavine globoko učenje mehanizem pozornosti river discharge precipitation deep learning attention mechanism true false false Slovenski jezik Angleški jezik Diplomsko delo/naloga 2025-09-15 11:15:06 2025-09-15 11:15:09 2025-11-07 03:51:48 0000-00-00 00:00:00 2025 0 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 38061 253392387 Furlan_Petja_-_Napovedovanje_recnega_pretoka_z_metodami_globokega_ucenja.pdf Furlan_Petja_-_Napovedovanje_recnega_pretoka_z_metodami_globokega_ucenja.pdf 1 B6F0EE74248327C6C6C3851DA6058348 1ea95490ccdd5dbf21fc3e125c4b544b0ab7efd0e418a12dfb04629ad9d11c2c 4ef3f886-9214-11f0-9328-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=217484 Fakulteta za računalništvo in informatiko 0 0 0