Zhiqiang Liao

Aalto University;

ISM Department;

Espoo, Finland.

I am currently a PhD student at the Department of Information and Service Management at Aalto University School of Business.

My research interests focus on Statistics, Optimization, and Statistical Machine Learning. I am also interested in their applications to regulation, finance, and sustainable development.

Before joining Aalto BIZ, I obtained my master degree in Business school at Sichuan University. My master thesis was entitled “Fuzzy support vector machine for preference learning and its application”.

Here you can find my publications page and my CV. Feel free to reach out.

Recent News

Sep 4, 2024	My first single-authored paper Variable selection in convex nonparametric least squares via structured Lasso: An application to the Swedish electricity market now is available from arXiv.
Apr 22, 2024	My paper Overfitting Reduction in Convex Regression co-authored with Sheng Dai, Eunji Lim, and Timo Kuosmanen now is available from arXiv.
May 5, 2023	My paper Convex Support Vector Regression co-authored with Sheng Dai and Timo Kuosmanen is accepted by European Journal of Operational Research.
Feb 14, 2023	I have updated my blogs Standardization in Machine Learning and Cross Validation in Machine Learning.
Oct 30, 2022	My paper Convex Support Vector Regression co-authored with Sheng Dai and Timo Kuosmanen now is available from arXiv.

Selected Publications

Convex support vector regression

Zhiqiang Liao, Sheng Dai, and Timo Kuosmanen

European Journal of Operational Research 2024

Abs HTML

Nonparametric regression subject to convexity or concavity constraints is increasingly popular in economics, finance, operations research, machine learning, and statistics. However, the conventional convex regression based on the least squares loss function often suffers from overfitting and outliers. This paper proposes to address these two issues by introducing the convex support vector regression (CSVR) method, which effectively combines the key elements of convex regression and support vector regression. Numerical experiments demonstrate the performance of CSVR in prediction accuracy and robustness that compares favorably with other state-of-the-art methods.