Imaging Time-Series to Improve Classification and Imputation

TL;DR

This paper introduces a novel image-based encoding of time series data to leverage computer vision techniques for improved classification and imputation, demonstrating competitive results on standard datasets.

Contribution

The paper proposes a new framework converting time series into images like GASF, GADF, and MTF, enabling the application of deep learning methods from computer vision to time series analysis.

Findings

Achieved competitive classification accuracy on 20 datasets.

Reduced imputation MSE by up to 48%.

Provided analysis of learned features and weights.

Abstract

Inspired by recent successes of deep learning in computer vision, we propose a novel framework for encoding time series as different types of images, namely, Gramian Angular Summation/Difference Fields (GASF/GADF) and Markov Transition Fields (MTF). This enables the use of techniques from computer vision for time series classification and imputation. We used Tiled Convolutional Neural Networks (tiled CNNs) on 20 standard datasets to learn high-level features from the individual and compound GASF-GADF-MTF images. Our approaches achieve highly competitive results when compared to nine of the current best time series classification approaches. Inspired by the bijection property of GASF on 0/1 rescaled data, we train Denoised Auto-encoders (DA) on the GASF images of four standard and one synthesized compound dataset. The imputation MSE on test data is reduced by 12.18%-48.02% when compared to using the raw data. An analysis of the features and weights learned via tiled CNNs and DAs explains why the approaches work.