Thesis Defense - İbrahim Edhem Sakarya (MSIE)

İbrahim Edhem Sakarya - M.Sc. Industrial Engineering

Assoc.Prof. Erkun Kundakcıoğlu - Advisor

AUC MAXIMIZATION FOR BINARY CLASSIFICATION
USING COMBINATORIAL BENDERS CUTS

Date: 11.06.2019

Time: 15:00

Location: AB2 G03

Thesis Committee: 

Assoc. Prof. Erhun Kundakcıoğlu,  Özyeğin University

Assoc. Prof. Okan Örsan Özener,  Özyeğin University

 Prof. Tonguç Ünlüyurt, Sabancı University

Abstract

The purpose of this study is to maximize the area under Receiver Operating Characteristic curve for binary classication problems using a scoring-based mixed integer linear programming formulation. We investigate exact approaches using a reformulation, combinatorial Benders cuts, and heuristic bounding methods. Our study
presents computational results on benchmark datasets and paves the way for future studies on scoring-based approaches.

Bio

Ibrahim Edhem SAKARYA graduated from C agrbey Anatolian High School in 2012. He received his B.S. degree in Industrial Engineering from Özyeğin University in June 2017. After graduation, he joined the Master of Science program in Industrial Engineering at Özyeğgin University and has been working under the supervision of Assoc.
Prof. Erhun Kundakcioglu. His research mainly focuses on combinatorial optimization and optimization in data mining.