Prediction framework in a distributed lag model with a target function: an application to global warming data

Özbay N., Toker Kutay S.

ENVIRONMENTAL AND ECOLOGICAL STATISTICS, cilt.28, sa.1, ss.87-134, 2021 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 28 Sayı: 1
  • Basım Tarihi: 2021
  • Doi Numarası: 10.1007/s10651-020-00477-x
  • Dergi Adı: ENVIRONMENTAL AND ECOLOGICAL STATISTICS
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, IBZ Online, ABI/INFORM, Aquatic Science & Fisheries Abstracts (ASFA), BIOSIS, CAB Abstracts, Environment Index, Geobase, Greenfile, Pollution Abstracts, Veterinary Science Database, Civil Engineering Abstracts
  • Sayfa Sayıları: ss.87-134
  • Anahtar Kelimeler: Distributed lag model, Global warming, Multicollinearity, Prediction mean square error, Target function
  • Çukurova Üniversitesi Adresli: Evet

Özet

Due to the nature of the distributed lag model, researchers are likely to encounter the problem of multicollinearity in this model. Biased estimation techniques, one of which is Almon ridge estimation, are alternatively considered instead of Almon estimation with the aim of recovering the multicollinearity. Although estimation performance is often taken into consideration, predictive performance is rarely handled in the distributed lag model. The principal purpose of this paper is to investigate the predictive performance of the distributed lag model through target function. In this context, we employ Almon ridge estimation to define a new predictor that is more resistant to multicollinearity. For an extensive analysis of the prediction problem in the distributed lag model, we concentrate on the theoretical results and comparisons. Then, the issue of determining optimal parameters is considered by means of minimizing the prediction mean square error. Numerical analysis depending on global warming data is examined to validate our theoretical outcomes. Moreover, a Monte Carlo experiment is carried out to evaluate the predictive ability of the estimators.