Learning from Irregularly-Sampled Time Series: A Missing Data Perspective
This addresses the challenge of handling irregularly-sampled time series for applications in healthcare and other domains, representing an incremental improvement over existing methods.
The paper tackles the problem of modeling irregularly-sampled time series, which are common in domains like healthcare, by treating them as missing data and using an encoder-decoder framework with methods like variational autoencoders and generative adversarial networks. The result is that their models achieve competitive or better classification results compared to recent RNN models while offering significantly faster training times.
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.