Earliness-Aware Deep Convolutional Networks for Early Time Series Classification

TL;DR

This paper introduces EA-ConvNets, a novel deep learning framework that jointly learns features and predicts early in time series data, outperforming existing methods and providing interpretable deep shapelets.

Contribution

The paper presents the first deep learning framework for data-driven feature learning in early time series classification, combining shapelet learning with dynamic truncation.

Findings

Outperforms state-of-the-art early classification methods

Achieves high accuracy on benchmark datasets

Provides interpretable deep shapelet features

Abstract

We present Earliness-Aware Deep Convolutional Networks (EA-ConvNets), an end-to-end deep learning framework, for early classification of time series data. Unlike most existing methods for early classification of time series data, that are designed to solve this problem under the assumption of the availability of a good set of pre-defined (often hand-crafted) features, our framework can jointly perform feature learning (by learning a deep hierarchy of \emph{shapelets} capturing the salient characteristics in each time series), along with a dynamic truncation model to help our deep feature learning architecture focus on the early parts of each time series. Consequently, our framework is able to make highly reliable early predictions, outperforming various state-of-the-art methods for early time series classification, while also being competitive when compared to the state-of-the-art time series classification algorithms that work with \emph{fully observed} time series data. To the best of our knowledge, the proposed framework is the first to perform data-driven (deep) feature learning in the context of early classification of time series data. We perform a comprehensive set of experiments, on several benchmark data sets, which demonstrate that our method yields significantly better predictions than various state-of-the-art methods designed for early time series classification. In addition to obtaining high accuracies, our experiments also show that the learned deep shapelets based features are also highly interpretable and can help gain better understanding of the underlying characteristics of time series data.