Shape-Adaptive Subspace-Constrained Quadratic Matrix Factorization
Organizations: Department of Statistics, Faculty of Arts and Sciences, Beijing Normal University, Zhuhai, China.
Abstract
Quadratic manifold fitting provides a flexible framework for modeling nonlinear low-dimensional structures embedded in high-dimensional spaces. However, existing methods typically employ prescribed noise models or fixed loss functions, limiting their adaptability to varying noise characteristics. To address this limitation, we propose an alternating maximum likelihood estimation method for subspace-constrained quadratic matrix factorization, termed AMLE-SQMF, which jointly estimates the quadratic manifold and the underlying noise distribution. Specifically, we model the residuals using a generalized Gaussian distribution (GGD), whose shape parameter controls the tail behavior and determines the associated fitting loss. By exploiting the closed-form maximum likelihood estimator of the scale parameter, we eliminate the scale variable and derive a profile negative log-likelihood. Based on this formulation, we develop an alternating optimization algorithm that iteratively updates the manifold parameters and estimates the noise shape parameter. This enables adaptive loss selection within a unified likelihood-based framework, without requiring prior specification of the noise characteristics. Experiments on synthetic and real-world datasets demonstrate the effectiveness of AMLE-SQMF in quadratic manifold fitting and its robustness to diverse noise distributions.
Figures & tables
| Methods / | 1.25 | 1.50 | 1.75 | 2.00 |
|---|---|---|---|---|
| GD | 13.91 | 3.01 | 12.41 | 0.81 |
| IRGU | 2.32 | 2.24 | 2.42 | 0.75 |
| 1.25 | 1.50 | 1.75 | 2.00 | |
|---|---|---|---|---|
| Mean | 0.046 | 0.032 | 0.033 | 0.036 |
| SD | 0.038 | 0.027 | 0.029 | 0.038 |
| = 0.05 | = 0.10 | = 0.15 | = 0.20 | |
|---|---|---|---|---|
| Shape | ||||
| 1.00 | 1.066 0.024 | 1.050 0.020 | 1.027 0.015 | 1.021 0.012 |
| 1.25 | 1.142 0.066 | 1.273 0.021 | 1.264 0.021 | 1.251 0.018 |
| 1.50 | 1.150 0.092 | 1.461 0.043 | 1.513 0.019 | 1.516 0.018 |
| 1.75 | 1.129 0.098 | 1.564 0.073 | 1.693 0.031 | 1.726 0.019 |
| 2.00 | 1.112 0.098 | 1.627 0.104 | 1.848 0.050 | 1.930 0.033 |
| = 30 | = 60 | = 90 | |
|---|---|---|---|
| Shape | |||
| 1.00 | 0.583 0.066 | 0.940 0.016 | 1.007 0.012 |
| 1.25 | 0.817 0.058 | 1.135 0.019 | 1.235 0.019 |
| 1.50 | 0.981 0.029 | 1.450 0.041 | 1.506 0.022 |
| 1.75 | 1.130 0.188 | 1.695 0.030 | 1.723 0.021 |
| 2.00 | 1.748 0.254 | 1.911 0.029 | 1.930 0.027 |
| Method | GGD = 1.25 | GGD = 1.50 | GGD = 1.75 | Logistic | Student- 10 |
|---|---|---|---|---|---|
| AMLE-SQMF (Ours) | 8.902 2.226 | 8.595 1.949 | 9.038 2.042 | 8.828 1.901 | 8.967 1.989 |
| SQMF ( = 1) | 9.857 2.465 | 10.443 4.573 | 10.247 1.971 | 9.642 1.676 | 9.756 1.551 |
| SQMF ( = 2) | 8.895 2.007 | 9.033 2.756 | 9.066 1.923 | 8.884 1.884 | 8.998 1.928 |
| LPCA | 12.121 2.637 | 12.049 2.333 | 12.163 2.532 | 11.815 2.394 | 11.914 2.299 |
| SPH | 14.889 5.137 | 14.457 4.230 | 14.565 4.823 | 14.583 4.272 | 14.485 3.832 |
| Q-LS | 11.923 2.399 | 11.825 2.357 | 11.816 2.251 | 11.614 2.162 | 11.660 2.137 |
| Zheng Zhai is currently a faculty member in the Department of Statistics, Faculty of Arts and Sciences at Beijing Normal University, Zhuhai. He received his Ph.D. degree from the School of Mathematical Sciences at Zhejiang University and worked as a postdoctoral researcher at the University of Toronto and the National University of Singapore. His current research interests include diffusion model, matrix computation, and manifold learning. |
| Zhuan Liang is currently a graduate student in the Department of Statistics, Faculty of Arts and Sciences at Beijing Normal University, Zhuhai. He received his B.S. degree from the School of Mathematics and Statistics at Shandong University. His current research interests include diffusion models and matrix computation. |
| Sujun Guo is currently a graduate student in the Department of Statistics, Faculty of Arts and Sciences, Beijing Normal University, Zhuhai, China. She received the B.S. degree from the School of Mathematics and Statistics, Northeast Normal University. Her research interests include statistical methods and their applications in machine learning. |
Manifold Learning
Manifold learning finds low-dimensional structure in complex data to support visualization, geometric analysis, and more efficient learning.
See all 124 papers