FIG is a method for dimensionality reduction and structure discovery in noisy, high-dimensional dynamical processes, with a focus on recovering latent geometry and enabling meaningful data visualization.
Overview of the FIG pipeline. Data are transformed into a functional basis, local covariance structure is estimated, functional PCA is performed, and distances are computed via a functional Mahalanobis metric.
The PHATE package is required.
There is a Python demo on how to use FIG.
Haozhe Chen, Andres Felipe Duque Correa, Guy Wolf, Kevin R. Moon
“Data Visualization using Functional Data Analysis,” SampTA 2025.
Preprint version:
https://arxiv.org/abs/2406.03396
If you use this code or build upon this work, please cite:
@inproceedings{chen2025fig,
title={Data Visualization using Functional Data Analysis},
author={Chen, Haozhe and Duque Correa, Andres Felipe and Wolf, Guy and Moon, Kevin R},
booktitle={Proceedings of SampTA},
year={2025}
}