Time series extrinsic regression (2020-06-23T00:00:00.000000Z)

TL;DR

The results show that the state-of-the-art TSC algorithm Rocket, when adapted for regression, achieves the highest overall accuracy compared to adaptations of other TSC algorithms and state of theart machine learning (ML) algorithms such as XGBoost, Random Forest and Support Vector Regression.

Abstract

This paper studies time series extrinsic regression (TSER): a regression task of which the aim is to learn the relationship between a time series and a continuous scalar variable; a task closely related to time series classification (TSC), which aims to learn the relationship between a time series and a categorical class label. This task generalizes time series forecasting, relaxing the requirement that the value predicted be a future value of the input series or primarily depend on more recent values. In this paper, we motivate and study this task, and benchmark existing solutions and adaptations of TSC algorithms on a novel archive of 19 TSER datasets which we have assembled. Our results show that the state-of-the-art TSC algorithm Rocket, when adapted for regression, achieves the highest overall accuracy compared to adaptations of other TSC algorithms and state-of-the-art machine learning (ML) algorithms such as XGBoost, Random Forest and Support Vector Regression. More importantly, we show that much research is needed in this field to improve the accuracy of ML models. We also find evidence that further research has excellent prospects of improving upon these straightforward baselines.

Authors

Geoffrey I. Webb

6 papers

F. Petitjean

3 papers

Chang Wei Tan

2 papers

TL;DR

Abstract

Authors

References72 items

Deep learning for time series classification

Monash University, UEA, UCR Time Series Regression Archive

Monash University, UEA, UCR Time Series Extrinsic Regression Archive

FFORMA: Feature-based forecast model averaging

FastEE: Fast Ensembles of Elastic Distances for time series classification

ROCKET: exceptionally fast and accurate time series classification using random convolutional kernels

InceptionTime: Finding AlexNet for time series classification

Deep PPG: Large-Scale Heart Rate Estimation with Convolutional Neural Networks

catch22: CAnonical Time-series CHaracteristics

Temporal Convolutional Neural Network for the Classification of Satellite Image Time Series

Transfer learning for time series classification

The UCR time series archive

The M4 Competition: Results, findings, conclusion and way forward

Deep learning for time series classification: a review

Proximity Forest: an effective and scalable distance-based classifier for time series

A fuel moisture content and flammability monitoring methodology for continental Australia based on optical remote sensing

Multi-Source Social Feedback of Online News Feeds

Multivariate Time Series Classification with WEASEL+MUSE

Computationally efficient heart rate estimation during physical exercise using photoplethysmographic signals

Methods for Scalar‐on‐Function Regression

Visualising forecasting algorithm performance using time series instance spaces

Fast and Accurate Time Series Classification with WEASEL

HIVE-COTE: The Hierarchical Vote Collective of Transformation-Based Ensembles for Time Series Classification

The great time series classification bake off: a review and experimental evaluation of recent algorithmic advances

Time series classification from scratch with deep neural networks: A strong baseline

Toward a Robust Estimation of Respiratory Rate From Pulse Oximeters

XGBoost: A Scalable Tree Boosting System

Generalizing DTW to the multi-dimensional case requires an adaptive approach

A Novel Time-Varying Spectral Filtering Algorithm for Reconstruction of Motion Artifact Corrupted Heart Rate Signals During Intense Physical Activities Using a Wearable Photoplethysmogram Sensor

The BOSS is concerned with time series classification in the presence of noise

Time-Series Classification with COTE: The Collective of Transformation-Based Ensembles

Time series classification with ensembles of elastic distance measures

Learning a symbolic representation for multivariate time series classification

Photoplethysmography-Based Heart Rate Monitoring in Physical Activities via Joint Sparse Spectrum Reconstruction

TROIKA: A General Framework for Heart Rate Monitoring Using Wrist-Type Photoplethysmographic Signals During Intensive Physical Exercise

Learning time-series shapelets

Estimator selection and combination in scalar-on-function regression

SAX-VSM: Interpretable Time Series Classification Using SAX and Vector Space Model

Highly comparative time-series analysis: the empirical structure of time series and their methods

A time series forest for classification and feature extraction

A shapelet transform for time series classification

Rotation-invariant similarity in time series using bag-of-patterns representation

Photoplethysmographic derivation of respiratory rate: a review of relevant physiology

Logical-shapelets: an expressive primitive for time series classification

Encyclopedia of Machine Learning

Scikit-learn: Machine Learning in Python

Time series shapelets: a new primitive for data mining

Forecasting with Exponential Smoothing: The State Space Approach

On field calibration of an electronic nose for benzene estimation in an urban pollution monitoring scenario

Statistical Comparisons of Classifiers over Multiple Data Sets

Machine Learning Benchmarks and Random Forest Regression

The M3-Competition: results, conclusions and implications

Support Vector Regression Machines

Support-Vector Networks

The accuracy of extrapolation (time series) methods: Results of a forecasting competition

Time Series Analysis: Forecasting and Control

The Holt-Winters Forecasting Procedure

A Comparison of Alternative Tests of Significance for the Problem of $m$ Rankings

The M4 Competition: 100,000 time series and 61 forecasting methods

Efficient search of the best warping window for Dynamic Time Warping

A brief history of time series forecasting competitions

UCImachinelearningrepository

Tree Boosting With XGBoost - Why Does XGBoost Win "Every" Machine Learning Competition?

Probabilistic Estimation of Respiratory Rate from Wearable Sensors

Fast Shapelets: A Scalable Algorithm for Discovering Time Series Shapelets

CapnoBase: Signal database and tools to collect, share and annotate respiratory signals

Forecasting with Exponential Smoothing

Random Forests

Grid-search optimised SVR

Exponential smoothing: The state of the art

Time series extrinsic regression

Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations

Field of Study

Journal Information

Name

Page