The Comparison of Robust Partial Least Squares Regression Methods (RSIMPLS, PRM) with Robust Principal Component Regression for Predicting Tourist Arrivals to Turkey

Esra Polat

Öz


Tourism is one of the most important component in the economic development strategy of many developing countries such as Turkey. The annual data set of Turkey (1986 - 2013), including the six factors affecting the tourist arrivals, is examined. The aim of this study is modelling the tourist arrivals to Turkey in cases of both multicollinearity and outlier existence in the data set by using a robust Principal Component Regression method: RPCR, two robust Partial Least Squares Regression methods: RSIMPLS and Partial Robust M-Regression (PRM). Hence, the best model giving the best predictions of tourist arrivals is selected and the most important factors are determined.


Anahtar Kelimeler


multicollinearity; outliers; robust principal component regression; robust partial least squares regression; tourist arrivals

Tam Metin:

PDF (English)


 

İndekslendiği Kaynaklar

 

Listeleyen Diğer Kaynaklar ve Servisler

 

Lisans

Creative Commons Lisansı
Doğuş Üniversitesi Dergisi'nin içeriği Creative Commons Atıf-Gayriticari 4.0 Uluslararası Lisansı ile lisanslanmıştır.
 

İletişim:

Doğuş Üniversitesi Dergisi
Acıbadem Zeamet Sokak, No: 21
34722 - Kadıköy, İSTANBUL
E-posta: journal@dogus.edu.tr