Prognoza vremenskih serija u senzorskim tokovima podataka

master's thesis

• Title (english): Time Series Forecasting in Sensor Data Streams
• Author: Antonio Ivčec
• Mentor: Krešimir Pripužić (mentor)
• Committee member: Krešimir Pripužić (predsjednik povjerenstva)
• Committee member: Dragan Jevtić (član povjerenstva)
• Committee member: Marin Vuković (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2018-07-05, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Kroz ovaj rad napravljen je pregled metoda analize i prognoze vremenskih serija. Objašnjene su metode prognoziranja vremenskih serija strojnim učenjem te strategije prognoze nekoliko koraka unaprijed. Detaljno su objašnjene metode prognoziranja neuronskim mrežama te statističkim autoregresivnim modelima. Na studijskom slučaju stvarnih senzorskih tokova vodostaja, protjecaja i padalina, implementirane su i evaluirane tehnike multivarijatne prognoze povratnim neuronskim mrežama te modelom autoregresije vektora. Izvršena je analiza nad senzorskim tokovima, a svi dobiveni rezultati su prikazani grafički.

Abstract (english)

In this thesis the methods of time series analysis and time series forecast are analyzed. Machine learning methods for forecasting and multi step strategies are described. Neural networks and statistical autoregressive models for forecasting are also described in detail. Recurrent neural network models and vector autoregression models are implemented and evaluated on the case study of actual sensory streams of water levels, sprains and rainfall. Analysis and forecast results are presented and visualized.

• Keywords: vremenske serije; senzorski tokovi; povratne neuronske mreže; autoregresivni model; vektorska autoregresija; analiza vremenskih serija; prognoza vremenskih serija
• Keywords (english): time series; sensory streams; recurrent neural networks; autoregressive model; vector autoregression; time series analysis; time series forecast
• Language: croatian
• URN:NBN: urn:nbn:hr:168:895852
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-13 21:01:57