Tien-En Chang (張天恩)
Tien-En Chang (張天恩)
Institute of Industrial Engineering, National Taiwan University
Embedded Files
I am Tien-En (Willy) Chang, a PhD candidate in Industrial Engineering at National Taiwan University (NTU), working with Prof. Argon Chen in the Statistical Data Mining Lab.
My research develops interpretable statistical and machine learning methods for feature attribution and variable selection. A central theme of my work is how to assign importance to variables when predictors are dependent, where standard methods can become ambiguous. I study computationally efficient feature importance methods, their theoretical properties under multicollinearity, and their connection to Shapley value attribution and model selection.
Beyond linear models, I am interested in extending these ideas toward modern explainable machine learning. During my internship at Polytechnique Montréal–Mila with Prof. Hervé Lombaert, I worked on lightweight and anatomically interpretable deep learning approaches for knee osteoarthritis assessment.
Recent News
Recent News
Mar. 2026: Our paper, "Variable Selection Using Relative Importance Rankings", has been accepted for publication in Pattern Recognition [paper].
Jan. 2026: Our paper, "Understanding and Using the Relative Importance Measures Based on Orthogonalization and Reallocation", has been accepted for publication in Statistical Science [paper].
Sep. 2025: Our work, “Anatomically-Focused Patches for Lightweight and Explainable Knee OA Grading,” received the Best Paper Award in MICCAI ShapeMI 2025 [paper].
Aug. 2025: One workshop paper accepted to ShapeMI 2025 in MICCAI 2025! See you at Daejeon!
Feb. 2025-Jun. 2025: Joined 2025 Polytechnique Montreal Winter Research Internship Program.
Research topic: Machine Learning and Interaction with Large-Scale Medical Imaging Datasets. Under the supervition of Prof. Herve Lombaert in the PolyShape Lab. Internship Program Report: 蒙特婁理工學院實習心得
May. 2024: One abstract accepted to DATA 2024! See you at Dijon!
Tien-En Chang, Argon Chen. Understanding the Orthonormality Transformation based Measures of Relative Importance.
Nov. 2023: Celebrating our team's success: Awarded the Outstanding Award at the Intelligent Manufacturing and Big Data Analytics Contest 2023.
The finals of Project B: the image classification problem in the manufacturing. Team name: 真替你們感到難過; members: Huai-Wei Wang, Zih-An Yi, Tien-En Chang; mentor: Jakey Blue
Oct. 2023: One abstract accepted to IASC-ARS 2023! See you at Sydney!
Tien-En Chang, Argon Chen. Using Relative Weight for Variable Importance Assessment - How Does It Work and Could It Work Better?
Page updated
Google Sites
Report abuse