Primjena metode potpornih vektora za dugoročno predviđanje potrošnje

master's thesis

• Title (english): Long term load forecasting based on support vector machine
• Author: Barbara Brna
• Mentor: Igor Kuzle (mentor)
• Committee member: Igor Kuzle (predsjednik povjerenstva)
• Committee member: Tomislav Tomiša (član povjerenstva)
• Committee member: Vladimir Mikuličić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2014-07-15, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

Dugoročno predviđanje potrošnje električne energije ima važnu ulogu prilikom planiranja izgradnje i obnavljanja infrastrukture. Sa napretkom tehnologije, umjetna inteligencija ima sve veću ulogu u predviđanju. Uz genetske algoritme, neuronske mreže, "fuzzy" logiku, "wavelet" mreže i ekspertne sustave, najvažnija metoda je metoda potpornih vektora. Ta metoda koristi strojno učenje za razvijanje modela kojim tada radi predviđanje. U ovom radu metoda potpornih vektora korištena je za dugoročno predviđanje potrošnje Krapinsko-zagorske županije. Faktori kojima je razvijen model su bili potrošnja u prošlosti, te BDP Krapinsko-zagorske županije. Prikazani su rezultati predviđanja dobiveni različitim metodama provjere, te je odabrano predviđanje koje daje najvjerojatnije rezultate.

Abstract (english)

Long term load forecasting has a very important role in developing and reconstruction of infrastructure. Because of technology advances, artificial intelligence is becoming an substantial tool in load forecasting. Besides genetic algorithms, neural networks, fuzzy logic, wavelet networks and expert systems, the most important method is Support Vector Machine. This method uses machine learning to develop a model, which is then used for forecasting. In this thesis support vector machines are used for long term load forecasting in Krapina-Zagorje County. The model is developed using past load demand data and gross domestic product in Krapina-Zagorje County. Finally, results are shown for different methods of validation, and the prediction with the most probable results has been chosen.

• Keywords: dugoročno predviđanje potrošnje; umjetna inteligencija; strojno učenje; metoda potpornih vektora; kernel; LS-SVM; objektno orijentirano sučelje
• Keywords (english): long term load forecasting; artificial intelligence; machine learning; support vector machine; kernel; LS-SVM; object oriented model interface
• Language: croatian
• URN:NBN: urn:nbn:hr:168:523601
• Study programme: Title: Electrical Engineering and Information Technology; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka elektrotehnike i informacijske tehnologije (magistar/magistra inženjer/inženjerka elektrotehnike i informacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2018-02-06 08:04:13