University of Sussex
Browse
Gjoreski_ISWC - camera ready.pdf (935.27 kB)

Unsupervised online activity discovery using temporal behaviour assumption

Download (935.27 kB)
conference contribution
posted on 2023-06-09, 07:13 authored by Hristijan GjoreskiHristijan Gjoreski, Daniel RoggenDaniel Roggen
We present a novel unsupervised approach, UnADevs, for discovering activity clusters corresponding to periodic and stationary activities in streaming sensor data. Such activities usually last for some time, which is exploited by our method; it includes mechanisms to regulate sensitivity to brief outliers and can discover multiple clusters overlapping in time to better deal with deviations from nominal behaviour. The method was evaluated on two activity datasets containing large number of activities (14 and 33 respectively) against online agglomerative clustering and DBSCAN. In a multi-criteria evaluation, our approach achieved significantly better performance on majority of the measures, with the advantages that: (i) it does not require to specify the number of clusters beforehand (it is open ended); (ii) it is online and can find clusters in real time; (iii) it has constant time complexity; (iv) and it is memory efficient as it does not keep the data samples in memory. Overall, it has managed to discover 616 of the total 717 activities. Because it discovers clusters of activities in real time, it is ideal to work alongside an active learning system.

Funding

Lifelearn: Unbounded activity and context awareness; G1786; EPSRC-ENGINEERING & PHYSICAL SCIENCES RESEARCH COUNCIL; EP/N007816/1

History

Publication status

  • Published

File Version

  • Accepted version

Journal

International Symposium on Wearable Computers

Publisher

Association for Computing Machinery

Issue

3

Volume

1

Page range

42-49

Event name

The 21st International Symposium on Wearable Computers (ISWC)

Event location

Maui, Hawai, USA

Event type

conference

Event date

11.09 - 15.09.2017

ISBN

9781450351881

Department affiliated with

  • Engineering and Design Publications

Research groups affiliated with

  • Sensor Technology Research Centre Publications

Full text available

  • Yes

Peer reviewed?

  • Yes

Legacy Posted Date

2017-07-17

First Open Access (FOA) Date

2017-10-17

First Compliant Deposit (FCD) Date

2017-07-14

Usage metrics

    University of Sussex (Publications)

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC