Skip to contents

fda.clust: Clustering Functional Data

Manuel Oviedo (UDC), Manuel Febrero (USC)

2024-12-16

Source: vignettes/Introduction.Rmd
Introduction.Rmd

Introduction

The fda.clust package provides specialized clustering methods for functional data. Inspired by the functional data analysis framework, it offers tools for:

  • Clustering functional data using functional k-means, DBSCAN, mean-shift, and hierarchical clustering.
  • Validating the quality of the resulting clusters using internal clustering measures.
  • Access to real-world datasets like ECG200, ECG5000, and growth_ldata.

Clustering Methods

The following clustering methods are provided:

  • fkmeans: Functional k-means clustering.
  • fdbscan: Functional DBSCAN clustering.
  • fmeanshift: Functional mean-shift clustering.
  • fhclust: Hierarchical clustering for functional data.

Datasets

Three datasets are included to demonstrate the use of clustering methods:

  • ECG200: Contains electrical signals from heartbeats, categorized into two groups: normal and myocardial infarction.
  • ECG5000: A larger dataset of heartbeats with 5,000 samples, representing heartbeats categorized into four groups.
  • growth_ldata: Longitudinal growth data from the Berkeley Growth Study, with height measurements of boys and girls.

More information about these datasets is available in their respective documentation.