Seyed Amir Hamed Zakeri, M.Sc.
Department of Statistics
November 2015, 81 pages

Supervisor: Ceylan Yozgatlıgil (Department of Statistics, Middle East Technical University, Ankara)
Co-supervisor:
Ömür Uğur (Institute of Applied Mathematics, Middle East Technical University, Ankara)

Abstract

Due to the liberalization of the electricity market, prices are now determined based on contracts on regulated markets and their behavior is mainly driven by constant supply and demand forces. Power producers and consumers need accurate price forecasting tools in a competitive market. Price forecasts give important information for producers and consumers to plan bidding strategies to maximize their benefits and utilities. Analysis of hourly electricity prices in Turkey is challenging due to the existence of multiple seasonality. In this study, we construct a time series model and obtain short-term forecasts of hourly electricity prices using multiple regression method. We used lagged price values, demand as the exogenous variable and dummy variables for Saturdays and Sundays to capture the seasonality in the price.

Keywords: Electricity Price Forecasting, Time Series, ARMAX, GARCH