Toward Personalized Activity Recognition Systems With a Semipopulation Approach
The DOI matches its Crossref record; 46 of the 47 references it cites were found in a registry or an index.
Abstract
Activity recognition is a key component of context-aware computing to support people's physical activity, but conventional approaches often lack in their generalizability and scalability due to problems of diversity in how individuals perform activities, overfitting when building activity models, and collection of a large amount of labeled data from end users. To address these limitations, we propose a semipopulation-based approach that exploits activity models trained from other users; therefore, a new user does not need to provide a large volume of labeled activity data. Instead of relying on any additional information from users like their weight or height, our approach directly measures the fitness of others' models on a small amount of labeled data collected from the new user. With these shared activity models among users, we compose a hybrid model of Bayesian networks and support vector machines to accurately recognize the activity of the new user. On activity data collected from 28 people with a diversity in gender, age, weight, and height, our approach produced an average accuracy of 83.4% (kappa: 0.852), compared with individual and (standard) population models that had accuracies of 77.3% (kappa: 0.79) and 77.7% (kappa: 0.743), respectively. Through an analysis on the performance of our approach and users' demographic information, our approach outperforms others that rely on users' demographic information for recognizing their activities, which may contradict the commonly held belief that physically similar people would have similar activity patterns.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“In contrast, real-world HAR systems may fail to distinct to new members and/or situations, resulting in low computational efficiency in practical applications [13], [14].”
From Deep CNN-LSTM With Self-Attention Model for Human Activity Recognition Using Wearable Sensor (2022) · cites this paper as [13]
In the citing paper
“While extensive explore has been completed on the effects of human traits on recognition accuracy, few studies have looked into the effects of position and device features [13], [15]–[18].”
From Deep CNN-LSTM With Self-Attention Model for Human Activity Recognition Using Wearable Sensor (2022) · cites this paper as [13]
Citations
Cites
A robust layered control system for a mobile robot
Brooks, R. (1986) · IEEE Journal on Robotics and Automation · doi:10.1109/jra.1986.1087032 · Cited by 7,799
Evaluation: from precision, recall and F-measure to ROC, informedness, markedness and correlation
Powers, D. M. W. (2020) · arXiv · doi:10.48550/arxiv.2010.16061 · Cited by 5,440
Neural Networks and the Bias/Variance Dilemma
Geman, S., Bienenstock, E., & Doursat, R. (1992) · Neural Computation · doi:10.1162/neco.1992.4.1.1 · Cited by 3,578 · cited 1×
“Both situations are to be avoided as they can contribute to poor performance [17].”
BackgroundActivity recognition using cell phone accelerometers
Kwapisz, J. R., Weiss, G. M., & Moore, S. A. (2011) · ACM SIGKDD Explorations Newsletter · doi:10.1145/1964897.1964918 · Cited by 2,912 · cited 1×
“In this study, we do not directly deal with the issue of sensor placement, but instead adopted a sensor placement approach commonly used [2], [3], [19], [21], [22], [24], [32], [40].”
MethodA Survey on Human Activity Recognition using Wearable Sensors
Lara, O. D., & Labrador, M. A. (2013) · IEEE Communications Surveys & Tutorials · doi:10.1109/surv.2012.110112.00192 · Cited by 2,662 · cited 6×
“In the activity recognition community, there is an on-going open debate about the relative merits of subject-dependent (or individual) models and subject-independent (or population) models [26].”
BackgroundInfluentialAssessing Agreement on Classification Tasks: The Kappa Statistic
Carletta, J. (1996) · arXiv · doi:10.48550/arxiv.cmp-lg/9602004 · Cited by 2,088
The Opportunity challenge: A benchmark database for on-body sensor-based activity recognition
Chavarriaga, R., Sagha, H., Calatroni, A., Digumarti, S. T., Tröster, G., Millán, J. D. R., & Roggen, D. (2013) · Pattern Recognition Letters · doi:10.1016/j.patrec.2012.12.014 · Cited by 742
A Knowledge-Driven Approach to Activity Recognition in Smart Homes
Chen, L., Nugent, C. D., & Wang, H. (2012) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2011.51 · Cited by 589
USC-HAD: a daily activity dataset for ubiquitous activity recognition using wearable sensors
Zhang, M., & Sawchuk, A. A. (2012) · Proceedings of the 2012 ACM Conference on Ubiquitous Computing · doi:10.1145/2370216.2370438 · Cited by 554 · cited 1×
“This is especially true for medical, military, and activities-of-daily-living-based applications [12], [15], [26], [27], [30], [43].”
BackgroundKhan, A. M., Lee, Y.-K., Lee, S. Y., & Kim, T.-S. (2010) · IEEE Transactions on Information Technology in Biomedicine · doi:10.1109/titb.2010.2051955 · Cited by 518 · cited 2×
“Advances in mobile sensing technology have led to the development of a number of human-activity recognition methods and systems, mostly using accelerometers [4], [11], [18], [22], [23], [47].”
BackgroundMethodFleury, A., Vacher, M., & Noury, N. (2010) · IEEE Transactions on Information Technology in Biomedicine · doi:10.1109/titb.2009.2037317 · Cited by 489 · cited 1×
“This is especially true for medical, military, and activities-of-daily-living-based applications [12], [15], [26], [27], [30], [43].”
BackgroundActivity recognition on streaming sensor data
Krishnan, N. C., & Cook, D. J. (2014) · Pervasive and Mobile Computing · doi:10.1016/j.pmcj.2012.07.003 · Cited by 488 · cited 1×
“Advances in mobile sensing technology have led to the development of a number of human-activity recognition methods and systems, mostly using accelerometers [4], [11], [18], [22], [23], [47].”
BackgroundSensor Positioning for Activity Recognition Using Wearable Accelerometers
Atallah, L., Lo, B., King, R., & Yang, G.-Z. (2011) · IEEE Transactions on Biomedical Circuits and Systems · doi:10.1109/tbcas.2011.2160540 · Cited by 346 · cited 2×
“First, our current system uses a number of wearable sensors, but users may only wear a few in practice [2], [3], [5], [21], [31].”
BackgroundMethodElderly activities recognition and classification for applications in assisted living
Chernbumroong, S., Cang, S., Atkins, A., & Yu, H. (2013) · Expert Systems with Applications · doi:10.1016/j.eswa.2012.09.004 · Cited by 317 · cited 1×
“This is especially true for medical, military, and activities-of-daily-living-based applications [12], [15], [26], [27], [30], [43].”
BackgroundCOSAR: hybrid reasoning for context-aware activity recognition
Riboni, D., & Bettini, C. (2011) · Personal and Ubiquitous Computing · doi:10.1007/s00779-010-0331-7 · Cited by 281 · cited 3×
“Riboni and Bettini exploited ontological reasoning combined with statistical inference [36].”
BackgroundMethodHuman Daily Activity Recognition With Sparse Representation Using Wearable Sensors
Zhang, M., & Sawchuk, A. A. (2013) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2013.2253613 · Cited by 184 · cited 5×
“In contrast, trying to model a user’s activity from a small amount of training data results in a limited ability to handle variations in how this user performs activities and is prone to overfitting [36], [44], [45].”
BackgroundMethodInfluentialActivity logging using lightweight classification techniques in mobile devices
Martín, H., Bernardos, A. M., Iglesias, J., & Casar, J. R. (2013) · Personal and Ubiquitous Computing · doi:10.1007/s00779-012-0515-4 · Cited by 180 · cited 1×
“First, our current system uses a number of wearable sensors, but users may only wear a few in practice [2], [3], [5], [21], [31].”
BackgroundLane, N. D., Xu, Y., Lu, H., Hu, S., Choudhury, T., Campbell, A. T., & Zhao, F. (2011) · Proceedings of the 13th international conference on Ubiquitous computing · doi:10.1145/2030112.2030160 · Cited by 171 · cited 2×
“[25] have shown that the physical conditions or lifestyles of users can be used to help identify activity (mobility) models for new users that share those characteristics.”
BackgroundA benchmark dataset to evaluate sensor displacement in activity recognition
Baños, O., Damas, M., Pomares, H., Rojas, I., Tóth, M. A., & Amft, O. (2012) · Proceedings of the 2012 ACM Conference on Ubiquitous Computing · doi:10.1145/2370216.2370437 · Cited by 163 · cited 2×
“First, our current system uses a number of wearable sensors, but users may only wear a few in practice [2], [3], [5], [21], [31].”
BackgroundMethodActiServ: Activity Recognition Service for mobile phones
Berchtold, M., Budde, M., Gordon, D., Schmidtke, H. R., & Beigl, M. (2010) · International Symposium on Wearable Computers (ISWC) 2010 · doi:10.1109/iswc.2010.5665868 · Cited by 160 · cited 1×
“Advances in mobile sensing technology have led to the development of a number of human-activity recognition methods and systems, mostly using accelerometers [4], [11], [18], [22], [23], [47].”
BackgroundPBN: towards practical activity recognition using smartphone-based body sensor networks
Keally, M., Zhou, G., Xing, G., Wu, J., & Pyles, A. (2011) · Proceedings of the 9th ACM Conference on Embedded Networked Sensor Systems · doi:10.1145/2070942.2070968 · Cited by 155 · cited 2×
“First, our current system uses a number of wearable sensors, but users may only wear a few in practice [2], [3], [5], [21], [31].”
BackgroundMethodWeakly Supervised Recognition of Daily Life Activities with Wearable Sensors
Stikic, M., Larlus, D., Ebert, S., & Schiele, B. (2011) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2011.36 · Cited by 155 · cited 3×
“Individual models typically rely on the collection and annotation of large amounts of user-specific training data, which are often very burdensome to end users and limit the acceptance of such activity recognition systems [37], [38].”
BackgroundMethodHuman motion recognition using a wireless sensor-based wearable system
Varkey, J. P., Pompili, D., & Walls, T. A. (2012) · Personal and Ubiquitous Computing · doi:10.1007/s00779-011-0455-4 · Cited by 144 · cited 1×
“In this study, we do not directly deal with the issue of sensor placement, but instead adopted a sensor placement approach commonly used [2], [3], [19], [21], [22], [24], [32], [40].”
MethodA Pattern Mining Approach to Sensor-Based Human Activity Recognition
Gu, T., Wang, L., Wu, Z., Tao, X., & Lu, J. (2011) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2010.184 · Cited by 139 · cited 1×
“Advances in mobile sensing technology have led to the development of a number of human-activity recognition methods and systems, mostly using accelerometers [4], [11], [18], [22], [23], [47].”
BackgroundHong, J.-H., Min, J.-K., Cho, U.-K., & Cho, S. (2007) · Pattern Recognition · doi:10.1016/j.patcog.2007.07.004 · Cited by 137 · cited 1×
“It then uses the SVMs, ordered according to their probabilities, in a subsumption architecture [7], [20], i.”
MethodSelf-citationCross-people mobile-phone based activity recognition
Zhao, Z., Chen, Y., Liu, J., Shen, Z., & Liu, M. (2011) · doi:10.5591/978-1-57735-516-8/ijcai11-423 · Cited by 134 · cited 1×
“, present a transfer learning technique in which the model of a new user is adapted from the model of an existing user [46].”
BackgroundMultimodal Physical Activity Recognition by Fusing Temporal and Cepstral Information
Li, M., Rozgic, V., Thatte, G., Lee, S., Emken, A., Annavaram, M., Mitra, U., Spruijt-Metz, D., & Narayanan, S. (2010) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2010.2053217 · Cited by 124 · cited 5×
“, in [14] and [27]; otherwise, the evaluation leads to overoptimistic performance results that will not hold when applied in the field [35].”
BackgroundMethodInfluentialUsing unlabeled data in a sparse-coding framework for human activity recognition
Bhattacharya, S., Nurmi, P., Hammerla, N., & Plötz, T. (2014) · Pervasive and Mobile Computing · doi:10.1016/j.pmcj.2014.05.006 · Cited by 121
A hierarchical approach to real-time activity recognition in body sensor networks
Wang, L., Gu, T., Tao, X., & Lü, J. (2010) · Pervasive and Mobile Computing · doi:10.1016/j.pmcj.2010.12.001 · Cited by 119
Motion- and location-based online human daily activity recognition
Zhu, C., & Sheng, W. (2010) · Pervasive and Mobile Computing · doi:10.1016/j.pmcj.2010.11.004 · Cited by 114
Complex activity recognition using context-driven activity theory and activity signatures
Saguna, S., Zaslavsky, A., & Chakraborty, D. (2013) · ACM Transactions on Computer-Human Interaction · doi:10.1145/2490832 · Cited by 103 · cited 3×
“Individual models typically rely on the collection and annotation of large amounts of user-specific training data, which are often very burdensome to end users and limit the acceptance of such activity recognition systems [37], [38].”
BackgroundPärkkä, J., Cluitmans, L., & Ermes, M. (2010) · IEEE Transactions on Information Technology in Biomedicine · doi:10.1109/titb.2010.2055060 · Cited by 86 · cited 1×
“In this study, we do not directly deal with the issue of sensor placement, but instead adopted a sensor placement approach commonly used [2], [3], [19], [21], [22], [24], [32], [40].”
MethodMagherini, T., Fantechi, A., Nugent, C., & Vicario, E. (2013) · IEEE Transactions on Human-Machine Systems · doi:10.1109/tsmc.2013.2283661 · Cited by 74 · cited 1×
“This is especially true for medical, military, and activities-of-daily-living-based applications [12], [15], [26], [27], [30], [43].”
BackgroundChavarriaga, R., Bayati, H., & Millán, J. D. R. (2011) · Personal and Ubiquitous Computing · doi:10.1007/s00779-011-0493-y · Cited by 55
Large-Scale Personalized Human Activity Recognition Using Online Multitask Learning
Sun, X., Kashima, H., & Ueda, N. (2013) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2012.246 · Cited by 54 · cited 3×
“Based on a variety of features investigated in other work [14], [26], [27], [39], we extract the mean and standard deviation of each channel.”
BackgroundMethodPersonalized mobile physical activity recognition
Reiss, A., & Stricker, D. (2013) · Proceedings of the 2013 International Symposium on Wearable Computers · doi:10.1145/2493988.2494349 · Cited by 53 · cited 1×
“and uses labeled data from a new user to learn weights for each classifier in a weighted majority vote [34].”
MethodA system for activity recognition using multi-sensor fusion
Gao, L., Bourke, A. K., & Nelson, J. (2011) · 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society · doi:10.1109/iembs.2011.6091939 · Cited by 45 · cited 2×
“Here, features, either orientation-free or calibrated constantly in orientation, are recommended to cope with sensor displacement since end users are often not consistent or welltrained in wearing such systems [7], [16].”
BackgroundMethodComparing Supervised Learning Techniques on the Task of Physical Activity Recognition
Dalton, A. F., & OLaighin, G. M. (2012) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/titb.2012.2223823 · Cited by 45 · cited 4×
“Based on a variety of features investigated in other work [14], [26], [27], [39], we extract the mean and standard deviation of each channel.”
BackgroundMethodInfluentialActivity recognition with hand-worn magnetic sensors
Maekawa, T., Kishino, Y., Sakurai, Y., & Suyama, T. (2012) · Personal and Ubiquitous Computing · doi:10.1007/s00779-012-0556-8 · Cited by 42 · cited 1×
“, introduced the concept of supportive users, in which they trained activity models by including a new user’s and others’ data [29].”
MethodRobust Human Activity and Sensor Location Corecognition via Sparse Signal Representation
Xu, W., Zhang, M., Sawchuk, A. A., & Sarrafzadeh, M. (2012) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2012.2211355 · Cited by 39 · cited 1×
“More sophisticated features can be incorporated such as correlation and entropy to improve recognition accuracy [14], [26], [27], [39], [42], [45].”
BackgroundLu, C.-H., Ho, Y.-C., Chen, Y., & Fu, L. (2013) · IEEE Transactions on Human-Machine Systems · doi:10.1109/tsmc.2013.2281586 · Cited by 33 · cited 2×
“To evaluate our semipopulation calibration approach, we use the leave-one subject/session-out cross validation [27], [28] that separates training and test data in terms of a data collection condition and compare our approaches with common approaches: hybrid versus nonhybrid, and semipopulation versus individual and po…”
BackgroundMethodTowards Robust Activity Recognition for Everyday Life: Methods and Evaluation
Reiss, A., Hendeby, G., & Stricker, D. (2013) · Proceedings of the ICTs for improving Patients Rehabilitation Research Techniques · doi:10.4108/icst.pervasivehealth.2013.251928 · Cited by 30 · cited 1×
“, in [14] and [27]; otherwise, the evaluation leads to overoptimistic performance results that will not hold when applied in the field [35].”
BackgroundZhang, S., McClean, S. I., & Scotney, B. W. (2012) · IEEE Transactions on Information Technology in Biomedicine · doi:10.1109/titb.2012.2188534 · Cited by 29 · cited 3×
“In contrast, trying to model a user’s activity from a small amount of training data results in a limited ability to handle variations in how this user performs activities and is prone to overfitting [36], [44], [45].”
BackgroundMethodBergmann, J., Graham, S., Howard, N., & McGregor, A. H. (2013) · Measurement · doi:10.1016/j.measurement.2013.03.004 · Cited by 8 · cited 1×
“First, our current system uses a number of wearable sensors, but users may only wear a few in practice [2], [3], [5], [21], [31].”
BackgroundEvaluating Human Activity Recognition Systems for AAL Environments
Álvarez-García, J. A. (2013) · Communications in computer and information science · doi:10.1007/978-3-642-37419-7_10 · Cited by 3 · cited 1×
“We chose these activities because they correspond to the most basic and common activities in people’s daily living and are useful for supporting assisted living environments for the elderly [1].”
BackgroundDarSens: a framework for distributed activity recognition from body-worn sensors
Haslgrübler, M., & Holzmann, C. (2010) · International Conference on Body Area Networks · doi:10.1145/2221924.2221969 · Cited by 3 · cited 1×
“In this study, we do not directly deal with the issue of sensor placement, but instead adopted a sensor placement approach commonly used [2], [3], [19], [21], [22], [24], [32], [40].”
Method
Research works printed but not linked: 1
- cvetkovic, 0, Semi-supervised learning for adaptation of human activity recognition classifier to the user, Proc Workshop Space Time Ambient Intell, 24
Cited by
Deep CNN-LSTM With Self-Attention Model for Human Activity Recognition Using Wearable Sensor
Khatun, M. A., Yousuf, M. A., Ahmed, S., Uddin, M. Z., Alyami, S. A., Al-Ashhab, S., Akhdar, H. F., Khan, A., Azad, A., & Moni, M. A. (2022) · IEEE Journal of Translational Engineering in Health and Medicine · doi:10.1109/jtehm.2022.3177710 · Cited by 288
“While extensive explore has been completed on the effects of human traits on recognition accuracy, few studies have looked into the effects of position and device features [13], [15]–[18].”
Robust Human Activity Recognition Using Smartphone Sensors via CT-PCA and Online SVM
Chen, Z., Zhu, Q., Soh, Y. C., & Zhang, L. (2017) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2017.2712746 · Cited by 247
“Since smartphones are widely used in our daily life and nonintrusive for users, they can be a good candidate for human activity recognition [8].”
BackgroundLonini, L., Dai, A., Shawen, N., Simuni, T., Poon, C., Shimanovich, L., Daeschler, M., Ghaffari, R., Rogers, J. A., & Jayaraman, A. (2018) · npj Digital Medicine · doi:10.1038/s41746-018-0071-z · Cited by 209
Distilling the Knowledge From Handcrafted Features for Human Activity Recognition
Chen, Z., Zhang, L., Cao, Z., & Guo, J. (2018) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2018.2789925 · Cited by 160
“Since they are widely available and nonintrusive for users, human activity recognition with smartphone sensors becomes more and more popular [10], [11].”
BackgroundOn the Personalization of Classification Models for Human Activity Recognition
Ferrari, A., Micucci, D., Mobilio, M., & Napoletano, P. (2020) · IEEE Access · doi:10.1109/access.2020.2973425 · Cited by 146
“However, real-world HAR systems may achieve non satisfying recognition accuracy in real world applications because HAR techniques may struggle to generalize to new users and/or new environments [14], [15].”
BackgroundMethodInfluential
Metadata
- Authors
- Hong, J.-H., Ramos, J., & Dey, A. K.
- Affiliations
- Samsung Electronics (South Korea) (KR)
- Carnegie Mellon University (US)
- Venue
- IEEE Transactions on Human-Machine Systems
- Published
- 2016-02
- Volume
- 46
- Issue
- 1
- Pages
- 101-112
- ISSN
- 2168-2291 (ISSN-L)
- 2168-2305
- OpenAlex
- W2285292066
- Cited by
- 94
- References
- 47
- Citations by year
- 2025: 6 · 2024: 3 · 2023: 11 · 2022: 13 · 2021: 11 · 2020: 16 · 2019: 4 · 2018: 12 · 2017: 11 · 2016: 7
- FWCI
- 8.67
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- IEEE Transactions on Human-Machine Systems (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Context-Aware Activity Recognition Systems
- Social Sciences › Transportation › Human Mobility and Location-Based Analysis
- Health Sciences › Medicine › Public Health, Environmental and Occupational Health › Physical Activity and Health
- Keywords
- activity recognition, context-aware computing, physical activity, Bayesian networks, support vector machine, hybrid model, demographic diversity, overfitting
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.