Skip to content

Research at St Andrews

Sensor-based human activity mining using Dirichlet process mixtures of directional statistical models

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We have witnessed an increasing number of activity-aware applications being deployed in real-world environments, including smart home and mobile healthcare. The key enabler to these applications is sensor-based human activity recognition; that is, recognising and analysing human daily activities from wearable and ambient sensors. With the power of machine learning we can recognise complex correlations between various types of sensor data and the activities being observed. However the challenges still remain: (1) they often rely on a large amount of labelled training data to build the model, and (2) they cannot dynamically adapt the model with emerging or changing activity patterns over time. To directly address these challenges, we propose a Bayesian nonparametric model, i.e. Dirichlet process mixture of conditionally independent von Mises Fisher models, to enable both unsupervised and semi-supervised dynamic learning of human activities. The Bayesian nonparametric model can dynamically adapt itself to the evolving activity patterns without human intervention and the learning results can be used to alleviate the annotation effort. We evaluate our approach against real-world, third-party smart home datasets, and demonstrate significant improvements over the state-of-the-art techniques in both unsupervised and supervised settings.
Close

Details

Original languageEnglish
Title of host publicationProceedings 2019 IEEE International Conference on Data Science and Advanced Analytics (DSAA 2019)
EditorsLisa Singh, Richard De Veaux, George Karypis, Francesco Bonchi, Jennifer Hill
PublisherIEEE Computer Society
Pages154-163
Number of pages10
ISBN (Electronic)9781728144931
ISBN (Print)9781728144948
DOIs
Publication statusPublished - 5 Oct 2019
Event6th IEEE International Conference on Data Science and Advanced Analytics (DSAA'19) - Washington DC, United States
Duration: 5 Oct 20198 Oct 2019
Conference number: 6
http://dsaa2019.dsaa.co/

Publication series

NameProceedings of the International Conference on Data Science and Advanced Analytics
ISSN (Print)2472-1573

Conference

Conference6th IEEE International Conference on Data Science and Advanced Analytics (DSAA'19)
Abbreviated titleDSAA 2019
CountryUnited States
CityWashington DC
Period5/10/198/10/19
Internet address

Discover related content
Find related publications, people, projects and more using interactive charts.

View graph of relations

Related by author

  1. Distributed self-monitoring sensor networks via Markov switching Dynamic Linear Models

    Fang, L., Ye, J. & Dobson, S. A., 16 Jun 2019, Proceedings 2019 IEEE 13th International Conference on Self-Adaptive and Self-Organizing Systems (SASO 2019). IEEE Computer Society, p. 33-42 8780572

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

  2. Discovery and recognition of unknown activities

    Ye, J., Fang, L. & Dobson, S. A., 12 Sep 2016, p. 783-792. 10 p.

    Research output: Contribution to conferencePaperpeer-review

  3. XLearn: learning activity labels across heterogeneous datasets

    Ye, J., Dobson, S. A. & Zambonelli, F., Jan 2020, In: ACM Transactions on Intelligent Systems and Technology. 11, 2, 28 p., 17.

    Research output: Contribution to journalArticlepeer-review

  4. Lifelong learning in sensor-based human activity recognition

    Ye, J., Dobson, S. A. & Zambonelli, F., 18 Nov 2019, In: IEEE Pervasive Computing. 18, 3, p. 49-58 10 p.

    Research output: Contribution to journalArticlepeer-review

ID: 260325128

Top