Electricity price forecasting using hybrid time series models

---

Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Fen Edebiyat Fakültesi, İstatistik Bölümü, Türkiye

Tezin Onay Tarihi: 2018

Öğrenci: BÜŞRA TAŞ

Danışman: CEYLAN YOZGATLIGİL

Özet:

Accurate forecasting of hourly electricity price is very important in a competitive market. Decision makers highly benefit from accurate forecasting. Because electricity cannot be stored, shocks to demand or supply affect the electricity prices. As a result, electricity prices show high volatility. Additionally, it may have multiple levels of seasonality. Therefore, forecasting with conventional methods is very difficult. In this study, hybrid models are constructed with Seasonal Autoregressive Integrated Moving Average (SARIMA), TBATS and Neural Network models for the analysis of hourly electricity prices in Turkey. Time series can contain both linear and nonlinear patterns. Thus, using a hybrid model can give better results in forecasting. Both linear and nonlinear parts of the time series can be modeled by this approach. While SARIMA model and TBATS model are used to capture the linear behavior of the electricity price series. Neural Network is used to model the nonlinearity in the series. Electricity demand is used as exogenous variable. Different combinations of hybrid models and individual models are compared in terms of forecasting performance. The results indicate that mostly hybrid models outperform the individual models in one-week ahead and one-day ahead forecasting.