Multi-feature Clustering of Step Data using Multivariate Functional Principal Component Analysis

TL;DR

This paper introduces a novel clustering method for high-dimensional, zero-inflated step data from wearable devices, combining feature extraction with multivariate functional PCA to improve clustering accuracy.

Contribution

It develops a new statistical approach that integrates multiple features of step data with multivariate functional PCA, enabling effective clustering where traditional methods fail.

Findings

Significant improvement in clustering quality demonstrated through simulations.

Effective application to real wearable device data.

Method outperforms classical clustering techniques on complex step data.

Abstract

This paper presents a new statistical method for clustering step data, a popular form of health record data easily obtained from wearable devices. Since step data are high-dimensional and zero-inflated, classical methods such as K-means and partitioning around medoid (PAM) cannot be applied directly. The proposed method is a novel combination of newly constructed variables that reflect the inherent features of step data, such as quantity, strength, and pattern, and a multivariate functional principal component analysis that can integrate all the features of the step data for clustering. The proposed method is implemented by applying a conventional clustering method such as K-means and PAM to the multivariate functional principal component scores obtained from these variables. Simulation studies and real data analysis demonstrate significant improvement in clustering quality.