Fast and Flexible ADMM Algorithms for Trend Filtering

Abstract

This paper presents a fast and robust algorithm for trend filtering, a recently developed nonparametric regression tool. It has been shown that, for estimating functions whose derivatives are of bounded variation, trend filtering achieves the minimax optimal error rate, while other popular methods like smoothing splines and kernels do not. Standing in the way of a more widespread practical adoption, however, is a lack of scalable and numerically stable algorithms for fitting trend filtering estimates. This paper presents a highly efficient, specialized ADMM routine for trend filtering. Our algorithm is competitive with the specialized interior point methods that are currently in use, and yet is far more numerically robust. Furthermore, the proposed ADMM implementation is very simple, and importantly, it is flexible enough to extend to many interesting related problems, such as sparse trend filtering and isotonic trend filtering. Software for our method is freely available, in both the C and R languages.

22 pages, 10 figures; published in Journal of Computational and Graphical Statistics, 2015

References in corpus (1)

• The Falling Factorial Basis and Its Statistical Applications

Cited by in corpus (16)

• Deep Learning for Short-Term Traffic Flow Prediction
• Modular proximal optimization for multidimensional total-variation regularization
• Short term prediction of demand for ride hailing services: A deep learning approach
• Sparse Regression with Multi-type Regularized Feature Modeling
• ADMM-Net for Communication Interference Removal in Stepped-Frequency Radar
• Multiscale spatial density smoothing: an application to large-scale radiological survey and anomaly detection
• Trend Filtering -- I. A Modern Statistical Tool for Time-Domain Astronomy and Astronomical Spectroscopy
• Distributed and Scalable Optimization for Robust Proton Treatment Planning
• Convex programming with fast proximal and linear operators
• On the Complexity of the Weighted Fused Lasso
• Trend Filtering -- II. Denoising Astronomical Signals with Varying Degrees of Smoothness
• P-splines with an l1 penalty for repeated measures
• Primal-Dual Active-Set Methods for Isotonic Regression and Trend Filtering
• Exponential Family Trend Filtering on Lattices
• Frank-Wolfe Methods with an Unbounded Feasible Region and Applications to Structured Learning
• An Algorithm for Graph-Fused Lasso Based on Graph Decomposition