|
Komparativna analiza algoritama mašinskog učenja za predviđanje broja aktivnih bunara u sistemima vodosnabdevanja
Comparative analysis of machine learning llgorithms for active well prediction in water works
Abstract
Specifičan domen ovog rada je predviđanje operativnih parametara u upravljanju vodosnabdevanjem, a fokus je na predviđanju neophodnog broja aktivnih bunara. Efikasno upravljanje brojem aktivnih bunara je ključno za optimizaciju resursa i održavanje stabilnosti u snabdevanju vodom, posebno u sistemima sa promenljivim potrošačkim zahtevima i sezonskim varijacijama. Ovaj rad istražuje primenu dva različita algoritma mašinskog učenja, K-Nearest Neighbors (KNN) i Decision Tree Regressor, za predviđanje broja aktivnih bunara na osnovu istorijskih podataka samog izvorišta. Performanse oba modela su ocenjene korišćenjem Mean Squared Error (MSE), Root Mean Squared Error (RMSE) i Rsquared (R2 ) metrika. Cilj rada je poređenje efikasnosti i tačnosti ovih algoritama u specifičnom domenu.
Sažetak
The specific domain of this study is the prediction of operational parameters in water supply management, with a focus on predicting the necessary number of active wells. Efficient management of the number of active wells is crucial for resource optimization and maintaining stability in water supply, especially in systems with variable consumer demands and seasonal variations. This study explores the application of two different machine learning algorithms, K-Nearest Neighbors (KNN) and Decision Tree Regressor, for predicting the number of active wells based on historical data from the water source. The performance of both models is evaluated using Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R-squared (R2 ) metrics. The aim of the study is to compare the effectiveness and accuracy of these algorithms in the specific domain.
|