Electricity price forecast using diferent methods

U ovom diplomskom radu bilo je potrebno pronaći različite izvore koji se bave metodama za prognozu cijene električne energije. Kao glavne metode za prognozu cijene električne energije u ovom radu su metoda umjetne neuronske mreže i metoda strojnog učenja. Svaka od metoda je detaljno analizirana i objašnjena, zatim je napravljena paralela između te dvije metode. Nakon toga je osmišljen model za simulaciju i definirane su njegove osnovne faze provedbe. Simulacijski dio je napravljen u programu „Orange“, te su analizirane dobivene predviđene cijene električne energije u obje metode, cijene su uspoređene međusobno i još sa stvarnim cijenama koje se nalaze na hrvatskoj burzi električne energije. Na temelju rezultata i teorijskog dijela izveden je zaključak o točnosti obje metode i o samoj problematici prognoze koja je na kraju uspješno odrađena.In this thesis it was required to find different sources that deal with methods for predicting the price of electricity. The main methods for predicting the price of electricity in this paper are the artificial neural network method and the machine learning method. Each of the methods was studied and explained in detail, then a parallel was made between the two methods. After that, a simulation model was designed and its basic implementation phases were defined. The simulation part was created in the "Orange" program, and the obtained predicted electricity prices in both methods were analyzed, the prices were compared with each other and with the actual prices on the Croatian electricity exchange. Based on the results and the theoretical part, a conclusion was made about the accuracy of both methods and the very problem of forecasting, which was eventually successfully completed