Deep Functional Factor Models: Forecasting High-Dimensional Functional Time Series via Bayesian Nonparametric Factorization

Abstract

This paper introduces the Deep Functional Factor Model (DF2M), a Bayesian nonparametric model designed for analysis of high-dimensional functional time series. DF2M is built upon the Indian Buffet Process and the multi-task Gaussian Process, incorporating a deep kernel function that captures non-Markovian and nonlinear temporal dynamics. Unlike many black-box deep learning models, DF2M offers an explainable approach to utilizing neural networks by constructing a factor model and integrating deep neural networks within the kernel function. Additionally, we develop a computationally efficient variational inference algorithm to infer DF2M. Empirical results from four real-world datasets demonstrate that DF2M provides better explainability and superior predictive accuracy compared to conventional deep learning models for high-dimensional functional time series.

Cite

Text

Liu et al. "Deep Functional Factor Models: Forecasting High-Dimensional Functional Time Series via Bayesian Nonparametric Factorization." International Conference on Machine Learning, 2024.

Markdown

[Liu et al. "Deep Functional Factor Models: Forecasting High-Dimensional Functional Time Series via Bayesian Nonparametric Factorization." International Conference on Machine Learning, 2024.](https://mlanthology.org/icml/2024/liu2024icml-deep/)

BibTeX

@inproceedings{liu2024icml-deep,
  title     = {{Deep Functional Factor Models: Forecasting High-Dimensional Functional Time Series via Bayesian Nonparametric Factorization}},
  author    = {Liu, Yirui and Qiao, Xinghao and Pei, Yulong and Wang, Liying},
  booktitle = {International Conference on Machine Learning},
  year      = {2024},
  pages     = {31709-31727},
  volume    = {235},
  url       = {https://mlanthology.org/icml/2024/liu2024icml-deep/}
}