CAFS : Cost-Aware Features Selection Method for Multimodal Stress Monitoring on Wearable Devices
(2022) In IEEE Transactions on Biomedical Engineering 69(3). p.1072-1084- Abstract
Objective: Today, stress monitoring on wearable devices is challenged by the tension between high-detection accuracy and battery lifetime driven by multimodal data acquisition and processing. Limited research has addressed the classification cost on multimodal wearable sensors, particularly when the features are cost-dependent. Thus, we design a Cost-Aware Feature Selection (CAFS) methodology that trades-off between prediction-power and energy-cost for multimodal stress monitoring. Methods: CAFS selects the most important features under different energy-constraints, which allows us to obtain energy-scalable stress monitoring models. We further propose a self-aware stress monitoring method that intelligently switches among the... (More)
Objective: Today, stress monitoring on wearable devices is challenged by the tension between high-detection accuracy and battery lifetime driven by multimodal data acquisition and processing. Limited research has addressed the classification cost on multimodal wearable sensors, particularly when the features are cost-dependent. Thus, we design a Cost-Aware Feature Selection (CAFS) methodology that trades-off between prediction-power and energy-cost for multimodal stress monitoring. Methods: CAFS selects the most important features under different energy-constraints, which allows us to obtain energy-scalable stress monitoring models. We further propose a self-aware stress monitoring method that intelligently switches among the energy-scalable models, reducing energy consumption. Results: Using CAFS methodology on experimental data and simulation, we reduce the energy-cost of the stress model designed without energy constraints up to 94.37%. We obtain 90.98% and 95.74% as the best accuracy and confidence values, respectively, on unseen data, outperforming state-of-the-art studies. Analyzing our interpretable and energy-scalable models, we showed that simple models using only heart rate (HR) or skin conductance level (SCL), confidently predict acute stress for HR>93.30BPM and non-stress for SCL< 6.42 μS, but, outside these values, a multimodal model using respiration and pulse wave's features is needed for confident classification. Our self-aware acute stress monitoring proposal saves 10x energy and provides 88.72% of accuracy on unseen data. Conclusion: We propose a comprehensive solution for the cost-aware acute stress monitoring design addressing the problem of selecting an optimized feature subset considering their cost-dependency and cost-constraints. Significant: Our design framework enables long-term and confident acute stress monitoring on wearable devices.
(Less)
- author
- Momeni, Niloofar
LU
; Valdés, Adriana Arza ; Rodrigues, João ; Sandi, Carmen and Atienza, David
- organization
- publishing date
- 2022-03-01
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Cost-aware machine learning, Cost-constraints feature selection, Low-power wearable devices, Stress monitoring
- in
- IEEE Transactions on Biomedical Engineering
- volume
- 69
- issue
- 3
- pages
- 13 pages
- publisher
- IEEE - Institute of Electrical and Electronics Engineers Inc.
- external identifiers
-
- scopus:85115713116
- pmid:34543185
- ISSN
- 0018-9294
- DOI
- 10.1109/TBME.2021.3113593
- language
- English
- LU publication?
- yes
- additional info
- Funding Information: Manuscript received March 31, 2021; revised July 6, 2021 and August 27, 2021; accepted September 7, 2021. Date of publication September 20, 2021; date of current version February 21, 2022. This work was supported in part by the NCCR Robotics through the Symbiotic Drone project, and Synapsy [51NF40-125759, -158776, and -185897], in part by the Swiss National Science Foundation (NSF) through the ML-Edge research (GA No. 200020182009/1) projects, and the grants [31003AB-135710, 176206], in part by the ONR-G Grant (N62909-20-1-2063), and in part by the Oak Foundation to CS. (Corresponding author: Niloofar Momeni.) Niloofar Momeni is with the Embedded Systems Laboratory, 1015 Lausanne, Switzerland, and also with the Department of Mathematical Statistics, Lund University, Lund 22100, Sweden (e-mail: niloofar. mni@gmail.com). Publisher Copyright: © 1964-2012 IEEE.
- id
- ae40aea8-3a17-4ed9-b975-c5029ce8e999
- date added to LUP
- 2022-12-30 11:00:18
- date last changed
- 2024-08-05 10:33:54
@article{ae40aea8-3a17-4ed9-b975-c5029ce8e999, abstract = {{<p>Objective: Today, stress monitoring on wearable devices is challenged by the tension between high-detection accuracy and battery lifetime driven by multimodal data acquisition and processing. Limited research has addressed the classification cost on multimodal wearable sensors, particularly when the features are cost-dependent. Thus, we design a Cost-Aware Feature Selection (CAFS) methodology that trades-off between prediction-power and energy-cost for multimodal stress monitoring. Methods: CAFS selects the most important features under different energy-constraints, which allows us to obtain energy-scalable stress monitoring models. We further propose a self-aware stress monitoring method that intelligently switches among the energy-scalable models, reducing energy consumption. Results: Using CAFS methodology on experimental data and simulation, we reduce the energy-cost of the stress model designed without energy constraints up to 94.37%. We obtain 90.98% and 95.74% as the best accuracy and confidence values, respectively, on unseen data, outperforming state-of-the-art studies. Analyzing our interpretable and energy-scalable models, we showed that simple models using only heart rate (HR) or skin conductance level (SCL), confidently predict acute stress for HR>93.30BPM and non-stress for SCL< 6.42 μS, but, outside these values, a multimodal model using respiration and pulse wave's features is needed for confident classification. Our self-aware acute stress monitoring proposal saves 10x energy and provides 88.72% of accuracy on unseen data. Conclusion: We propose a comprehensive solution for the cost-aware acute stress monitoring design addressing the problem of selecting an optimized feature subset considering their cost-dependency and cost-constraints. Significant: Our design framework enables long-term and confident acute stress monitoring on wearable devices.</p>}}, author = {{Momeni, Niloofar and Valdés, Adriana Arza and Rodrigues, João and Sandi, Carmen and Atienza, David}}, issn = {{0018-9294}}, keywords = {{Cost-aware machine learning; Cost-constraints feature selection; Low-power wearable devices; Stress monitoring}}, language = {{eng}}, month = {{03}}, number = {{3}}, pages = {{1072--1084}}, publisher = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}}, series = {{IEEE Transactions on Biomedical Engineering}}, title = {{CAFS : Cost-Aware Features Selection Method for Multimodal Stress Monitoring on Wearable Devices}}, url = {{http://dx.doi.org/10.1109/TBME.2021.3113593}}, doi = {{10.1109/TBME.2021.3113593}}, volume = {{69}}, year = {{2022}}, }