Open Access Repository

Going through directional changes: evolving human movement classifiers using an event based encoding

Lones, MA, Alty, J ORCID: 0000-0002-5456-8676, Duggan-Carter, P, Turner, AJ, Jamieson, DRS and Smith, SL 2017 , 'Going through directional changes: evolving human movement classifiers using an event based encoding', in GECCO '17: Proceedings of the Genetic and Evolutionary Computation Conference Companion, Berlin, Germany 15-19 July , Association for Computing Machinery, New York, NY, United States, pp. 1365-1371 , doi: 10.1145/3067695.3082490.

Full text not available from this repository.

Abstract

Directional changes (DC) is an event based encoding for time seriesdata that has become popular in fnancial analysis, particularlywithin the evolutionary algorithm community. In this paper, we apply DC to a medical analytics problem, using it to identify and summarise the periods of opposing directional trends present within aset of accelerometry time series recordings. The summarised timeseries data are then used to train classifiers that can discriminatebetween different kinds of movement. As a case study, we consider the problem of discriminating the movements of Parkinson’sdisease patients when they are experiencing a common effect ofmedication called levodopa-induced dyskinesia. Our results suggest that a DC encoding is competitive against the window-basedsegmentation and frequency domain encodings that are often usedwhen solving this kind of problem, but offers added benefits in theform of faster training and increased interpretability.CCS CONCEPTS•Computing methodologies → Genetic programming; •Appliedcomputing → Health informatics;

Item Type: Conference Publication
Authors/Creators:Lones, MA and Alty, J and Duggan-Carter, P and Turner, AJ and Jamieson, DRS and Smith, SL
Keywords: bioengineering, Parkinson's disease, sensors, diagnostics, genetic programming, directional changes, time series analysis, movement analysis, Parkinson’s disease, dyskinesia
Publisher: Association for Computing Machinery
DOI / ID Number: 10.1145/3067695.3082490
Copyright Information:

Copyright 2017 ACM

Related URLs:
Item Statistics: View statistics for this item

Actions (login required)

Item Control Page Item Control Page
TOP