A short tutorial for time series classification and explanation with MrSQM

Thach Le Nguyen,Georgiana Ifrim

doi:10.1016/j.simpa.2021.100197

Abstract

This paper presents MrSQM, a Python tool for the task of time series classification and explanation. Time series classification is a critical problem not only in scientific research but also in many real-life applications. However, state-of-the-art time series classifiers including deep learning and ensemble architectures are often impractical due to their complexity. MrSQM can provide an alternative lightweight solution, just as accurate but faster, and explainable. The tool is written mainly in C++ but wrapped with Cython to provide a more accessible Python interface.

Highlights

A time series is a sequence of numerical data values collected over a period of time or based on some other ordering of values such as spatial ordering
Time series classification is the problem of assigning a class to an unseen time series
The tool is written in C++ and Python

Summary

Introduction

A time series is a sequence of numerical data values collected over a period of time (e.g., the number of steps a person takes every minute [1]) or based on some other ordering of values such as spatial ordering (e.g., the shape of a coffee leaf or historical artefact [2]). A time series classifier is typically deemed useful when it is (1) accurate, (2) efficient, and (3) explainable. The visualization using a saliency map (Fig. 1) highlights the important parts of the time series with regard to the prediction. Its performance is comparable to state-of-the-art time series classifiers (e.g., Inception Time [9], TS-CHIEF [10], HIVE-COTE [11], ROCKET [12]). It can provide a saliency map to explain the classification prediction by highlighting the parts of the time series that most influenced the classification decision

Methodology

Implementation

Example

Software impact