Learning from Irregularly-Sampled Time Series: A Missing Data Perspective

TL;DR

This paper presents a novel approach to modeling irregularly-sampled time series by treating them as missing data, introducing an encoder-decoder framework with continuous convolutional layers, and demonstrating improved efficiency and competitive accuracy.

Contribution

It introduces a new framework for irregular time series using variational autoencoders and GANs, with continuous convolutions for better modeling and faster training.

Findings

Achieves competitive or superior classification accuracy.

Offers significantly faster training times.

Effective modeling of irregularly-sampled multivariate time series.

Abstract

Irregularly-sampled time series occur in many domains including healthcare. They can be challenging to model because they do not naturally yield a fixed-dimensional representation as required by many standard machine learning models. In this paper, we consider irregular sampling from the perspective of missing data. We model observed irregularly-sampled time series data as a sequence of index-value pairs sampled from a continuous but unobserved function. We introduce an encoder-decoder framework for learning from such generic indexed sequences. We propose learning methods for this framework based on variational autoencoders and generative adversarial networks. For continuous irregularly-sampled time series, we introduce continuous convolutional layers that can efficiently interface with existing neural network architectures. Experiments show that our models are able to achieve competitive or better classification results on irregularly-sampled multivariate time series compared to recent RNN models while offering significantly faster training times.