A Saddle Point Algorithm for Robust Data-Driven Factor Model Problems

2025/06/11 by Khodakaramzadeh, Shabnam, Shafiee, Soroosh, Gleizer, Gabriel de Albuquerque +1
#FOS: Electrical engineering #FOS: Mathematics #Optimization and Control (math.OC) #Systems and Control (eess.SY) #electronic engineering #information engineering

paper · doi:10.48550/arxiv.2506.09776

Abstract

We study the factor model problem, which aims to uncover low-dimensional structures in high-dimensional datasets. Adopting a robust data-driven approach, we formulate the problem as a saddle-point optimization. Our primary contribution is a general first-order algorithm that solves this reformulation by leveraging a linear minimization oracle (LMO). We further develop semi-closed form solutions (up to a scalar) for three specific LMOs, corresponding to the Frobenius norm, Kullback-Leibler divergence, and Gelbrich (aka Wasserstein) distance. The analysis includes explicit quantification of these LMOs' regularity conditions, notably the Lipschitz constants of the dual function, whthich govern the algorithm's convergence performance. Numerical experiments confirm our meod's effectiveness in high-dimensional settings, outperforming standard off-the-shelf optimization solvers.

Citations

• A Robust Optimization Approach to Network Control Using Local Information Exchange
• Hidden Factor estimation in Dynamic Generalized Factor Analysis Models
• A Robust Approach to ARMA Factor Modeling
• Bridging Bayesian and Minimax Mean Square Error Estimation via Wasserstein Distributionally Robust Optimization
• An alternating minimization algorithm for Factor Analysis
• Distributionally Robust Inverse Covariance Estimation: The Wasserstein\n Shrinkage Estimator
• Computation of the Maximum Likelihood estimator in low-rank Factor\n Analysis
• Factor Models with Real Data: a Robust Estimation of the Number of\n Factors
• Dykstra's Algorithm, ADMM, and Coordinate Descent: Connections, Insights, and Extensions
• Factor Analysis of Moving Average Processes
• Proximal alternating linearized minimization for nonconvex and nonsmooth problems
• Modeling complex systems by Generalized Factor Analysis
• Stochastic Gradient Descent for Non-smooth Optimization: Convergence\n Results and Optimal Averaging Schemes
• Factor modeling for high-dimensional time series: Inference for the number of factors
• Large Covariance Estimation by Thresholding Principal Orthogonal\n Complements
• Guaranteed Minimum-Rank Solutions of Linear Matrix Equations via Nuclear Norm Minimization
• Some Matrix Rearrangement Inequalities
• The Generalized Dynamic-Factor Model: Identification and Estimation
• Decision-Theoretic Planning: Structural Assumptions and Computational Leverage
• The EM Algorithm and Extensions.
• "General Intelligence," Objectively Determined and Measured
• Compressed sensing
• A class of Wasserstein metrics for probability distributions.

Related

• A Newton Interior-Point Method for ℓ₀ Factor Analysis
• T-Rex: Fitting a Robust Factor Model via Expectation-Maximization
• Computation of the Maximum Likelihood estimator in low-rank Factor\n Analysis
• Factor Models with Real Data: a Robust Estimation of the Number of\n Factors
• Fitting Multilevel Factor Models