Thesis Defense - Zeynel Batuhan Organ (MSIE)

Zeynel Batuhan Organ – M.Sc. Industrial Engineering

Assist. Prof. Enis Kayış – Advisor

Assist. Prof. Tagi Hanalioğlu – Co-Advisor

Date: 30.06.2022

Time: 09.30

Location: AB1 - 231

“Rolling Look-Ahead Approaches for Optimal Classification Trees”

Assist. Prof. Enis Kayış, Özyeğin University

Assist. Prof. Tagi Hanalioğlu, Bilkent University

Prof. Mehmet Güray Güler, Yıldız Technical University

Assist. Prof. Dilek Günneç Danış, Özyeğin University

Assist. Prof. Erinç Albey, Özyeğin University

Abstract:

Classification trees have gained tremendous attention in machine learning applications due to their inherently interpretable nature. Current state-of-the-art formulations for learning optimal binary classification trees suffer from scalability for larger depths or larger instances. Moreover, they mostly fail to prove optimality after long run times and fit perfectly to the training data while minimizing misclassification error which is likely fail to generalize to the test data. We present a simple but powerful new formulation which we call the rolling look-ahead learning approach. By dropping tractability variables that are dependent on instance size, we present a novel 2-depth optimal binary classification tree formulation with the objective to minimize gini impurity or misclassification error. The approach can be thought of as a middle ground between myopic and global optimization methods. For larger depths, we developed a hybrid approach that learns by looking ahead 2-steps rolling horizon. It is much faster than the fastest known global optimization methods which can solve an instance with around 50K rows & 135 features in less than 4 minutes, for depth 8. Also, in the majority of cases, the proposed approach outperforms global optimization methods & CART in terms of win count tested for 7 depths, 10 Fold and 19 benchmark datasets, and increase in out-of-sample accuracy up to 16.8% and 11.9% with respect to global optimization methods and CART, respectively.

Bio:

Zeynel Batuhan Organ graduated from Özyeğin University in August 2018 with a B.Sc. in Industrial Engineering. He has been pursuing his M.Sc. degree in Industrial Engineering at Özyeğin University since September 2019, under the supervision of Dr. Enis Kayış and Dr. Taghi Khaniyev. His interests are optimal classification trees.