Comparing Energy Demand Estimation Using Various Statistical Methods: The Case of Turkey


BULUT Y. M., YILDIZ Z.

GAZI UNIVERSITY JOURNAL OF SCIENCE, cilt.29, sa.2, ss.237-244, 2016 (ESCI) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 29 Sayı: 2
  • Basım Tarihi: 2016
  • Dergi Adı: GAZI UNIVERSITY JOURNAL OF SCIENCE
  • Derginin Tarandığı İndeksler: Emerging Sources Citation Index (ESCI), Scopus
  • Sayfa Sayıları: ss.237-244
  • Anahtar Kelimeler: Energy Demand, Energy Modeling, Biased Regression, PARTIAL LEAST-SQUARES, GENETIC ALGORITHM APPROACH, EASTERN SAUDI-ARABIA, CONSUMPTION, REGRESSION, OPTIMIZATION, NETWORKS, MODEL
  • Eskişehir Osmangazi Üniversitesi Adresli: Evet

Özet

Many engineers and scientists concern with future energy demand. They use many different statistical methods to estimate future energy demand such as multiple linear regression, neural networks, genetic algorithms and so on. In this paper, we propose ridge regression (RR) and partial least squares regression (PLSR) methods to estimate future energy demand. Because of the fact that variables, which are used in energy demand, are very collinear, ridge regression and partial least squares regression methods give more realistic results than least squares regression method. So, energy demand equations are developed based on RR and PLSR methods. Since, RR give better estimation, we estimate Turkey's future energy demand based on RR method.