Bilkent University
Department of Computer Engineering

---

S E M I N A R

 

AUC-Based Discretization Algorithm

 

Murat Kurtcephe
MSc. Student
Computer Engineering Department
Bilkent University

We present a new discretization method based on Area Under Roc Curve (AUC) measure combined with Quick Hull algorithm. Our method is a global, static and supervised discretization method. The proposed method will be compared with Ent-MDLP which is known as one of the best Discretization methods, in terms of average weighted AUC of naïve Bayes and voting feature intervals (VFI) algorithms by using real world datasets. The empirical results shows that our method can be a good alternative to Entropy based discretization. Keywords: Discretization, AUC, QuickHull, Naive Bayes, VFI

 

DATE: 2 November, 2009, Monday @ 16:20
PLACE: EA 409