Unexpected properties of bandwidth choice when smoothing discrete data for constructing a functional data classifier
From MaRDI portal
Publication:2443201
Abstract: The data functions that are studied in the course of functional data analysis are assembled from discrete data, and the level of smoothing that is used is generally that which is appropriate for accurate approximation of the conceptually smooth functions that were not actually observed. Existing literature shows that this approach is effective, and even optimal, when using functional data methods for prediction or hypothesis testing. However, in the present paper we show that this approach is not effective in classification problems. There a useful rule of thumb is that undersmoothing is often desirable, but there are several surprising qualifications to that approach. First, the effect of smoothing the training data can be more significant than that of smoothing the new data set to be classified; second, undersmoothing is not always the right approach, and in fact in some cases using a relatively large bandwidth can be more effective; and third, these perverse results are the consequence of very unusual properties of error rates, expressed as functions of smoothing parameters. For example, the orders of magnitude of optimal smoothing parameter choices depend on the signs and sizes of terms in an expansion of error rate, and those signs and sizes can vary dramatically from one setting to another, even for the same classifier.
Recommendations
- scientific article; zbMATH DE number 2104211
- Optimal classification and nonparametric regression for functional data
- Bandwidth choice for nonparametric classification
- Two-sample tests in functional data analysis starting from discrete data
- Choosing trajectory and data type when classifying functional data
Cites work
- An Effective Bandwidth Selector for Local Least Squares Regression
- Bandwidth choice for nonparametric classification
- Bayesian Curve Classification Using Wavelets
- Componentwise classification and clustering of functional data
- Confidence bands for Horvitz-Thompson estimators using sampled noisy functional data
- Deciding the dimension of effective dimension reduction space for functional and high-dimension\-al data
- Discriminant and cluster analysis for Gaussian stationary processes: local linear fitting approach
- Functional Classification in Hilbert Spaces
- Functional Classification with Margin Conditions
- Functional data analysis.
- Functional Logistic Discrimination Via Regularized Basis Expansions
- Horvitz-Thompson estimators for functional data: asymptotic confidence bands and optimal allocation for stratified sampling
- scientific article; zbMATH DE number 991833 (Why is no real title available?)
- Introduction to Information Retrieval
- On Properties of Functional Principal Components Analysis
- Response-Adaptive Regression for Longitudinal Data
- Robust estimation and classification for functional data via projection-based depth notions
- Second-order comparison of Gaussian random functions and the geometry of DNA minicircles
- Theory for high-order bounds in functional principal components analysis
- Two-sample tests in functional data analysis starting from discrete data
- Unexpected properties of bandwidth choice when smoothing discrete data for constructing a functional data classifier
- Uniform convergence rates for nonparametric regression and principal component analysis in functional/longitudinal data
Cited in
(11)- Clustering multivariate functional data using unsupervised binary trees
- In-sample forecasting applied to reserving and mesothelioma mortality
- Unexpected properties of bandwidth choice when smoothing discrete data for constructing a functional data classifier
- Optimal classification and nonparametric regression for functional data
- A centroid classifier for functional/longitudinal data
- NONPARAMETRIC DENSITY ESTIMATION BY B-SPLINE DUALITY
- On functional logistic regression: some conceptual issues
- Symmetrisation of a class of two-sample tests by mutually considering depth ranks including functional spaces
- Nonparametric functional data classification and bandwidth selection in the presence of MNAR class variables
- Adaptive functional principal components analysis
- Nonparametric density estimation and bandwidth selection with B-spline bases: a novel Galerkin method
This page was built for publication: Unexpected properties of bandwidth choice when smoothing discrete data for constructing a functional data classifier
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2443201)
