Variation Pattern Classification of Functional Data

Shuhao Jiao,Ron D. Frostig,Hernando Ombao

DOI: https://doi.org/10.48550/arXiv.2004.00855

2022-02-05

Abstract:A new classification method for functional data is proposed in this paper. This work is motivated by the need to identify features that discriminate between neurological conditions on which local field potentials (LFPs) were recorded. Regardless of the condition, these local field potentials have zero mean and thus the first moments of these random processes do not have discriminating power. We propose the variation pattern classification (VPC) method {which employs the (auto-)covariance operators as the discriminating features} and uses the Hilbert-Schmidt norm to measure the discrepancy between the (auto-)covariance operators of different groups. The proposed VPC method is demonstrated to be sensitive to the discrepancy, {potentially leading to a higher rate of classification}. One important innovation lies in the dimension reduction where the VPC method data-adaptively determines the basis functions (discriminative feature functions) that account for the major discrepancy. In addition, the selected discriminative feature functions provide insights on the discrepancy between different groups because they reveal the features of variation pattern that differentiate groups. Consistency properties are established and, furthermore, simulation studies and the analysis of rat brain LFP trajectories empirically demonstrate the advantages and effectiveness of the proposed method.

Methodology

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to effectively identify the features that can distinguish different groups of functional data in functional data classification when the functional data of different groups have similar mean functions. Specifically, the paper focuses on how to distinguish pre - stroke and post - stroke brain signals through local field potential (LFPs) signals in simulated stroke experiments. Since LFPs fluctuate around zero regardless of the clinical condition, relying solely on the first - order moment (i.e., the mean) cannot provide sufficient discriminative power. The paper proposes a variability pattern classification (VPC) method based on (auto) - covariance operators. This method uses the Hilbert - Schmidt norm to measure the differences between (auto) - covariance operators among different groups and adaptively selects basis functions through data to capture the main differences, thereby improving the classification accuracy. The main contributions of the paper are as follows: 1. **Proposing a new classification method**: The VPC method can effectively distinguish different groups when the functional data of different groups have similar mean functions. 2. **Data - adaptive selection of basis functions**: The VPC method can adaptively select basis functions that can capture the maximum differences. These basis functions not only improve the classification accuracy but also reveal the characteristics of the differences between different groups. 3. **Considering the information within the curve**: The VPC method considers the information within the curve, which provides important information for distinguishing different groups. 4. **Applicable to independent and dependent functions**: The proposed framework can be applied to independent and dependent functions, expanding the application range of the method. Through these innovations, the paper aims to provide clinicians with a method for quickly detecting stroke onset so that treatment measures can be taken as early as possible to reduce the impact of stroke on patients.

Variation Pattern Classification of Functional Data

MULTIVARIATE VARYING COEFFICIENT MODEL FOR FUNCTIONAL RESPONSES

High Dimensional Classification for Spatially Dependent Data with Application to Neuroimaging

Interpretable principal component analysis for multilevel multivariate functional data

Optimal Linear Discriminant Analysis for High-Dimensional Functional Data

Analysis of nonlinear modes of variation for functional data

Interpretable Principal Components Analysis for Multilevel Multivariate Functional Data, with Application to EEG Experiments

A Sparse Representation-Based Algorithm For Pattern Localization In Brain Imaging Data Analysis

Filtrated Common Functional Principal Components for Multivariate Functional data

Interpretable discriminant analysis for functional data supported on random nonlinear domains with an application to Alzheimer's disease

A Method of Feature Selection via Deep Convolution Neural Networks For Encoding Nonlinear Functional Network Connectivity and Its Application To The Classification of Mental Disorders

Multilevel hybrid principal components analysis for region-referenced functional electroencephalography data

Pattern Classification of Visual Evoked Potentials Based on Parallel Factor Analysis

Modeling Spectral Properties in Stationary Processes of Varying Dimensions with Applications to Brain Local Field Potential Signals

Dynamic functional connections analysis with spectral learning for brain disorder detection

An MVPA Method Based on Sparse Representation for Pattern Localization in Fmri Data Analysis

Multivariate Deep Learning Classification of Alzheimer’s Disease Based on Hierarchical Partner Matching Independent Component Analysis

Break Point Detection for Functional Covariance

Sparse Functional Principal Component Analysis in High Dimensions

Functional Data Analysis for Sparse Longitudinal Data

Groupwise structural sparsity for discriminative voxels identification