TU Darmstadt / ULB / TUbiblio

Data augmentation for time series: traditional vs generative models on capacitive proximity time series

Fu, Biying and Kirchbuchner, Florian and Kuijper, Arjan (2020):
Data augmentation for time series: traditional vs generative models on capacitive proximity time series.
p. 10, ACM, 13th ACM International Conference on PErvasive Technologies Related to Assistive Environments (PETRA'20), Corfu, Greece, June 2020, ISBN 978-1-4503-7773-7,
DOI: 10.1145/3389189.3392606,
[Conference or Workshop Item]

Abstract

Large labeled quantities and diversities of training data are often needed for supervised, data-based modelling. Data distribution should cover a rich representation to support the generalizability of the trained end-to-end inference model. However, this is often hindered by limited labeled data and the expensive data collection process, especially for human activity recognition tasks. Extensive manual labeling is required. Data augmentation is thus a widely used regularization method for deep learning, especially applied on image data to increase the classification accuracy. But it is less researched for time series. In this paper, we investigate the data augmentation task on continuous capacitive time series with the example on exercise recognition. We show that the traditional data augmentation can enrich the source distribution and thus make the trained inference model more generalized. This further increases the recognition performance for unseen target data around 21.4 percentage points compared to inference model without data augmentation. The generative models such as variational autoencoder or conditional variational autoencoder can further reduce the variance on the target data.

Item Type: Conference or Workshop Item
Erschienen: 2020
Creators: Fu, Biying and Kirchbuchner, Florian and Kuijper, Arjan
Title: Data augmentation for time series: traditional vs generative models on capacitive proximity time series
Language: English
Abstract:

Large labeled quantities and diversities of training data are often needed for supervised, data-based modelling. Data distribution should cover a rich representation to support the generalizability of the trained end-to-end inference model. However, this is often hindered by limited labeled data and the expensive data collection process, especially for human activity recognition tasks. Extensive manual labeling is required. Data augmentation is thus a widely used regularization method for deep learning, especially applied on image data to increase the classification accuracy. But it is less researched for time series. In this paper, we investigate the data augmentation task on continuous capacitive time series with the example on exercise recognition. We show that the traditional data augmentation can enrich the source distribution and thus make the trained inference model more generalized. This further increases the recognition performance for unseen target data around 21.4 percentage points compared to inference model without data augmentation. The generative models such as variational autoencoder or conditional variational autoencoder can further reduce the variance on the target data.

Publisher: ACM
ISBN: 978-1-4503-7773-7
Uncontrolled Keywords: Activity recognition, Ambient intelligence (AmI), Sensor data exploration
Divisions: 20 Department of Computer Science
20 Department of Computer Science > Mathematical and Applied Visual Computing
Event Title: 13th ACM International Conference on PErvasive Technologies Related to Assistive Environments (PETRA'20)
Event Location: Corfu, Greece
Event Dates: June 2020
Date Deposited: 26 Oct 2020 12:16
DOI: 10.1145/3389189.3392606
Official URL: https://doi.org/10.1145/3389189.3392606
Export:
Suche nach Titel in: TUfind oder in Google
Send an inquiry Send an inquiry

Options (only for editors)
Show editorial Details Show editorial Details