Kernel Additive Principal Components

11/21/2015
by   Xin Lu Tan, et al.
0

Additive principal components (APCs for short) are a nonlinear generalization of linear principal components. We focus on smallest APCs to describe additive nonlinear constraints that are approximately satisfied by the data. Thus APCs fit data with implicit equations that treat the variables symmetrically, as opposed to regression analyses which fit data with explicit equations that treat the data asymmetrically by singling out a response variable. We propose a regularized data-analytic procedure for APC estimation using kernel methods. In contrast to existing approaches to APCs that are based on regularization through subspace restriction, kernel methods achieve regularization through shrinkage and therefore grant distinctive flexibility in APC estimation by allowing the use of infinite-dimensional functions spaces for searching APC transformation while retaining computational feasibility. To connect population APCs and kernelized finite-sample APCs, we study kernelized population APCs and their associated eigenproblems, which eventually lead to the establishment of consistency of the estimated APCs. Lastly, we discuss an iterative algorithm for computing kernelized finite-sample APCs.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
07/03/2021

Novel Semi-parametric Tobit Additive Regression Models

Regression method has been widely used to explore relationship between d...
research
05/25/2021

Kernel Knockoffs Selection for Nonparametric Additive Models

Thanks to its fine balance between model flexibility and interpretabilit...
research
03/29/2020

Nonparametric additive factor models using sieve methods

This paper proposes a nonparametric additive factor model where the comm...
research
07/12/2022

Envelopes and principal component regression

Envelope methods offer targeted dimension reduction for various models. ...
research
10/07/2019

Nonparametric principal subspace regression

In scientific applications, multivariate observations often come in tand...
research
03/20/2018

Data Distillery: Effective Dimension Estimation via Penalized Probabilistic PCA

The paper tackles the unsupervised estimation of the effective dimension...

Please sign up or login with your details

Forgot password? Click here to reset