Selecting Ridge Parameters in Infinite Dimensional Hypothesis Spaces

نویسندگان

  • Masashi Sugiyama
  • Klaus-Robert Müller
چکیده

Previously, an unbiased estimator of the generalization error called the subspace information criterion (SIC) was proposed for a finite dimensional reproducing kernel Hilbert space (RKHS). In this paper, we extend SIC so that it can be applied to any RKHSs including infinite dimensional ones. Computer simulations show that the extended SIC works well in ridge parameter selection.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Dimensionality Reduction and Learning : Ridge Regression vs . PCA

1 Intro The theme of these two lectures is that for L2 methods we need not work in infinite dimensional spaces. In particular, we can unadaptively find and work in a low dimensional space and achieve about as good results. These results question the need for explicitly working in infinite (or high) dimensional spaces for L2 methods. In contrast, for sparsity based methods (including L1 regulari...

متن کامل

Dispersion of Torsional Surface Wave in a Pre-Stressed Heterogeneous Layer Sandwiched Between Anisotropic Porous Half-Spaces Under Gravity

The study of surface waves in a layered media has their viable application in geophysical prospecting. This paper presents an analytical study on the dispersion of torsional surface wave in a pre-stressed heterogeneous layer sandwiched between a pre-stressed anisotropic porous semi-infinite medium and gravitating anisotropic porous half-space. The non-homogeneity within the intermediate layer a...

متن کامل

Infinite-dimensional versions of the primary, cyclic and Jordan decompositions

The famous primary and cyclic decomposition theorems along with the tightly related rational and Jordan canonical forms are extended to linear spaces of infinite dimensions with counterexamples showing the scope of extensions.

متن کامل

An extension theorem for finite positive measures on surfaces of finite‎ ‎dimensional unit balls in Hilbert spaces

A consistency criteria is given for a certain class of finite positive measures on the surfaces of the finite dimensional unit balls in a real separable Hilbert space. It is proved, through a Kolmogorov type existence theorem, that the class induces a unique positive measure on the surface of the unit ball in the Hilbert space. As an application, this will naturally accomplish the work of Kante...

متن کامل

Asymptotic properties of Lasso+mLS and Lasso+Ridge in sparse high-dimensional linear regression

Abstract: We study the asymptotic properties of Lasso+mLS and Lasso+ Ridge under the sparse high-dimensional linear regression model: Lasso selecting predictors and then modified Least Squares (mLS) or Ridge estimating their coefficients. First, we propose a valid inference procedure for parameter estimation based on parametric residual bootstrap after Lasso+ mLS and Lasso+Ridge. Second, we der...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2002