On preserving statistical characteristics of accelerometry data using their empirical cumulative distribution

Nils Y. Hammerla, Reuben Kirkham, Peter Andras, Thomas Pl̈otz

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

77 Citations (Scopus)

Abstract

The majority of activity recognition systems in wearable computing rely on a set of statistical measures, such as means and moments, extracted from short frames of continuous sensor measurements to perform recognition. These features implicitly quantify the distribution of data observed in each frame. However, feature selection remains challenging and labour intensive, rendering a more generic method to quantify distributions in accelerometer data much desired. In this paper we present the ECDF representation, a novel approach to preserve characteristics of arbitrary distributions for feature extraction, which is particularly suitable for embedded applications. In extensive experiments on six publicly available datasets we demonstrate that it outperforms common approaches to feature extraction across a wide variety of tasks.

Original languageEnglish
Title of host publicationISWC 2013 - Proceedings of the 2013 ACM International Symposium on Wearable Computers
Pages65-68
Number of pages4
DOIs
Publication statusPublished - 15 Oct 2013
Externally publishedYes
Event2013 17th ACM International Symposium on Wearable Computers, ISWC 2013 - Zurich, Switzerland
Duration: 9 Sep 201312 Sep 2013

Publication series

NameISWC 2013 - Proceedings of the 2013 ACM International Symposium on Wearable Computers

Conference

Conference2013 17th ACM International Symposium on Wearable Computers, ISWC 2013
Country/TerritorySwitzerland
CityZurich
Period9/09/1312/09/13

Keywords

  • Accelerometry
  • Activity recognition
  • Feature representation

Cite this