HAKKINDA

  • ISSN: 2149-4916
  • E-ISSN: 2149-9373
  • Yayın Periyodu:
    Nisan - Ağustos - Aralık
  • Editör:İsmail ŞAHİN
  • Yıl: 2017
  • Cilt: 3
  • No: 3

A STUDY ON CORRECT CLASSİFİCATİON PERFORMANCE OF UNCONSTRAİNT PARTİAL PROPORTİONAL ODDS MODEL

Hatice DAĞLIOĞLU,  Semra ORAL ERBAŞ, 

Several methods are used to classify units in terms of categorical variables having two or more levels. Some of these methods are proportional odds model which is used when the level of the dependent variable is more than two and a ordinal pattern and the unconstraint partial proportional odds model which is used when the fundamental assumption of this model is established from the some of the variables and is not satisfied some of the variables. In this study, the correct classification rates are investigated with varying the number of classes of dependent variables, the number of independent variables and the sample size by taking unconstraint partial proportional odds model. A simulation study is performed in the case of all independent variables are continuous and also superiority of proportional odds model, unconstraint partial proportional odds model, linear discriminant analyze and quadratic discriminant analyze to each other are tried to revealed for these data.

Anahtar Kelime: Proportional odds model,  unconstraint partial proportional odds model,  correct classification rate,  simulation

Tam Metin (PDF) Kaynakça


Makale Gönder