Detecting and removing outlier(s) in electromyographic gait-related patterns

Miler Jerkovic Vera, Bojanic Dubravka, Jorgovanovic Nikola, Ilic Vojin and Petrovacki-Balj Bojana

Journal of Applied Statistics, 2013, vol. 40, issue 6, 1319-1332

Abstract: In this paper, we propose a method for outlier detection and removal in electromyographic gait-related patterns (EMG-GRPs). The goal was to detect and remove EMG-GRPs that reduce the quality of gait data while preserving natural biological variations in EMG-GRPs. The proposed procedure consists of general statistical tests and is simple to use. The Friedman test with multiple comparisons was used to find particular EMG-GRPs that are extremely different from others. Next, outlying observations were calculated for each suspected stride waveform by applying the generalized extreme studentized deviate test. To complete the analysis, we applied different outlier criteria. The results suggest that an EMG-GRP is an outlier if it differs from at least 50% of the other stride waveforms and contains at least 20% of the outlying observations. The EMG signal remains a realistic representation of muscle activity and demonstrates step-by-step variability once the outliers, as defined here, are removed.

Date: 2013
References: View complete reference list from CitEc
Citations:

Downloads: (external link)
http://hdl.handle.net/10.1080/02664763.2013.785495 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:japsta:v:40:y:2013:i:6:p:1319-1332

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/CJAS20

DOI: 10.1080/02664763.2013.785495

Access Statistics for this article

Journal of Applied Statistics is currently edited by Robert Aykroyd

More articles in Journal of Applied Statistics from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().