Abstract
The identification of daily activities is mainly limited by the imbalanced number of actions and the diversity of categories, which can lead to unsatisfactory classification results and difficult annotation. In this paper, a novel model is proposed to efficiently solve the mentioned problems. The model fully considers the characteristics of sample imbalance and uses the combination of Synthetic Minority Oversampling Technique and K-means clustering undersampling to achieve data balancing. After that, Active Learning is applied to query the most representative samples for labeling to reduce data annotation. Experiments show that the proposed method can balance the number of samples without increasing the computational cost of the system. Compared with the original data, the proposed model can learn the features of each class more fully after the samples are balanced, increasing the F1-score by 23%, and it can reduce the sample annotations by nearly 30% without reducing the recognition results.
Similar content being viewed by others
Data availability
The datasets generated during the current study are available in the [ECML SPHERE Challenge] repository, [https://doi.org/10.5523/bris.8gccwpx47rav19vk8x4xapcog].
References
Adaimi R, Thomaz E (2019) Leveraging active learning and conditional mutual information to minimize data annotation in human activity recognition [J]. Proc ACM Interact Mob Wearable Ubiquitous Technol 3(3):1–23. https://doi.org/10.1145/3351228
Akbari A, Castilla RS, Jafari R et al (2020) Using intelligent personal annotations to improve human activity recognition for movements in natural environments [J]. IEEE J Biomed Health Inform 24(9):2639–2650. https://doi.org/10.1109/JBHI.2020.2966151
Alani AA, Cosma G, Taherkhani A (2020) Classifying imbalanced multi-modal sensor data for human activity recognition in a smart home using deep learning [C]. In: 2020 international joint conference on neural networks (IJCNN). IEEE, pp 1–8. https://doi.org/10.1109/IJCNN48605.2020.9207697
Arafat MY, Hoque S, Farid DM (2017) Cluster-based under-sampling with random forest for multi-class imbalanced classification [C]. In: 2017 11th international conference on software, knowledge, information management and applications (SKIMA). IEEE, pp 1–6. https://doi.org/10.1109/SKIMA.2017.8294105
Arafat MY, Hoque S, Xu S et al (2019) An under-sampling method with support vectors in multi-class imbalanced data classification [C]. In: 2019 13th international conference on software, knowledge, information management and applications (SKIMA). IEEE, pp 1–6. https://doi.org/10.1109/SKIMA47702.2019.8982391
Bengar JZ, van de Weijer J, Twardowski B et al (2021) Reducing label effort: Self-supervised meets active learning [C]. In: Proceedings of the IEEE/CVF international conference on computer vision, pp 1631–1639. https://doi.org/10.1109/ICCVW54120.2021.00188
Bengar JZ, van de Weijer J, Fuentes LL et al (2022) Class-balanced active learning for image classification [C]. Proceedings of the IEEE/CVF winter conference on applications of computer vision, pp 1536–1545. https://doi.org/10.48550/arXiv.2110.04543
Bi H, Perello-Nieto M, Santos-Rodriguez R et al (2020) Human activity recognition based on dynamic active learning [J]. IEEE J Biomed Health Inform 25(4):922–934. https://doi.org/10.1109/JBHI.2020.3013403
Buda M, Maki A, Mazurowski MA (2018) A systematic study of the class imbalance problem in convolutional neural networks [J]. Neural Netw 106:249–259. https://doi.org/10.1016/j.neunet.2018.07.011
Bulling A, Blanke U, Schiele B (2014) A tutorial on human activity recognition using body-worn inertial sensors [J]. ACM Comput Surv (CSUR) 46(3):1–33. https://doi.org/10.1145/2499621
Chawla NV, Bowyer KW, Hall LO et al (2002) SMOTE: synthetic minority over-sampling technique [J]. J Artif Intell Res 16:321–357. https://doi.org/10.1613/jair.953
Chen Z, Jiang C, Xiang S et al (2019) Smartphone sensor-based human activity recognition using feature fusion and maximum full a posteriori [J]. IEEE Trans Instrum Meas 69(7):3992–4001. https://doi.org/10.1109/TIM.2019.2945467
Chen K, Zhang D, Yao L et al (2021) Deep learning for sensor-based human activity recognition: overview, challenges, and opportunities [J]. ACM Comput Surv (CSUR) 54(4):1–40. https://doi.org/10.1145/3447744
Choi J, Yi KM, Kim J et al (2021) Vab-al: Incorporating class imbalance and difficulty with variational bayes for active learning [C]. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp 6749–6758. https://doi.org/10.48550/arXiv.2003.11249
Cipolla E, Infantino I, Maniscalco U et al (2017) Indoor actions classification through long short term memory neural networks [C]. In: International conference on image analysis and processing. Springer, Cham, pp 435–444. https://doi.org/10.1007/978-3-319-68560-1_39
Elsts A, Twomey N, McConville R et al (2020) Energy-efficient activity recognition framework using wearable accelerometers [J]. J Netw Comput Appl 168:s. https://doi.org/10.1016/j.jnca.2020.102770
Gao G, Li Z, Huan Z et al (2021) Human behavior recognition model based on feature and classifier selection [J]. Sensors 21(23):7791. https://doi.org/10.3390/s21237791
Guo S, Liu Y, Chen R et al (2019) Improved SMOTE algorithm to deal with imbalanced activity classes in smart homes [J]. Neural Process Lett 50(2):1503–1526. https://doi.org/10.1007/s11063-018-9940-3
Hamad RA, Kimura M, Lundström J (2020) Efficacy of imbalanced data handling methods on deep learning for smart homes environments [J]. SN Comput Sci 1:1–10. https://doi.org/10.1007/s42979-020-00211-1
Ho SS, Wechsler H (2008) Query by transduction [J]. IEEE Trans Pattern Anal Mach Intell 30(9):1557–1571. https://doi.org/10.1109/TPAMI.2007.70811
Huan Z, Wei C, Li GH (2018) Outlier detection in wireless sensor networks using model selection-based support vector data descriptions [J]. Sensors 18(12):4328. https://doi.org/10.3390/s18124328
Huan Z, Chen X, Lv S et al (2019) Gait recognition of acceleration sensor for smart phone based on multiple classifier fusion [J]. Math Probl Eng 2019. https://doi.org/10.1155/2019/6471532
Huan Z, Lv S, Hou Z et al (2020) An evaluation strategy for the symmetry and consistency of lower limb segments during upper limb loading [J]. IEEE Sensors J 21(5):6440–6449. https://doi.org/10.1109/JSEN.2020.3039865
Ijaz MF, Alfian G, Syafrudin M et al (2018) Hybrid prediction model for type 2 diabetes and hypertension using DBSCAN-based outlier detection, synthetic minority over sampling technique (SMOTE), and random forest [J]. Appl Sci 8(8):1325. https://doi.org/10.3390/app8081325
Ijaz MF, Attique M, Son Y (2020) Data-driven cervical cancer prediction model with outlier detection and over-sampling methods [J]. Sensors 20(10):2809. https://doi.org/10.3390/s20102809
Janicka M, Lango M, Stefanowski J (2019) Using information on class interrelations to improve classification of multiclass imbalanced data: a new resampling algorithm [J]. Int J Appl Math Comput Sci 29(4). https://doi.org/10.2478/amcs-2019-0057
Koziarski M (2020) Radial-based undersampling for imbalanced data classification [J]. Pattern Recogn 102:107262. https://doi.org/10.1016/j.patcog.2020.107262
Koziarski M, Woźniak M, Krawczyk B (2020) Combined cleaning and resampling algorithm for multi-class imbalanced data with label noise [J]. Knowl-Based Syst 204:106223. https://doi.org/10.1016/j.knosys.2020.106223
Kumar Y, Koul A, Singla R et al (2022) Artificial intelligence in disease diagnosis: a systematic literature review, synthesizing framework and future research agenda [J]. J Ambient Intell Humaniz Comput:1–28. https://doi.org/10.1007/s12652-021-03612-z
Kwon H, Abowd GD, Plötz T (2019) Handling annotation uncertainty in human activity recognition [C]. In: Proceedings of the 23rd international symposium on wearable computers, pp 109–117. https://doi.org/10.1145/3341163.3347744
Lewis DD, Gale WA (1994) A sequential algorithm for training text classifiers [C]//SIGIR’94. Springer, London, pp 3–12. https://doi.org/10.1145/219587.219592
Liu M, Dong M, Jing C (2021) A modified real-value negative selection detector-based oversampling approach for multiclass imbalance problems [J]. Inf Sci 556:160–176. https://doi.org/10.1016/j.ins.2020.12.058
Malki Z, Atlam E, Dagnew G et al (2020) Bidirectional residual LSTM-based human activity recognition [J]. Comput Inf Sci 13(3):40. https://doi.org/10.5539/cis.v13n3p40
Narasimman G, Lu K, Raja A et al A* HAR: a new benchmark towards semi-supervised learning for class-imbalanced human activity recognition [J]. arXiv preprint arXiv:2101.04859, 2021. https://doi.org/10.48550/arXiv.2101.04859
Nekooeimehr I, Lai-Yuen SK (2016) Adaptive semi-unsupervised weighted oversampling (A-SUWO) for imbalanced datasets [J]. Expert Syst Appl 46:405–416. https://doi.org/10.1016/j.eswa.2015.10.031
Oh S, Ashiquzzaman A, Lee D et al (2021) Study on human activity recognition using semi-supervised active transfer learning [J]. Sensors 21(8):2760. https://doi.org/10.3390/s21082760
Pruengkarn R, Wong KW, Fung CC (2017) Multiclass imbalanced classification using fuzzy C-mean and SMOTE with fuzzy support vector machine [C] International conference on neural information processing. Springer, Cham: 67–75. https://doi.org/10.1007/978-3-319-70139-4_7
Ramanujam E, Perumal T, Padmavathi S (2021) Human activity recognition with smartphone and wearable sensors using deep learning techniques: a review [J]. IEEE Sensors J. https://doi.org/10.1109/JSEN.2021.3069927
Song J, Huang X, Qin S et al (2016) A bi-directional sampling based on K-means method for imbalance text classification [C]. In: 2016 IEEE/ACIS 15th international conference on computer and information science (ICIS). IEEE, pp 1–5. https://doi.org/10.1109/ICIS.2016.7550920
Srinivasu PN, SivaSai JG, Ijaz MF et al (2021) Classification of skin disease using deep learning neural networks with MobileNet V2 and LSTM [J]. Sensors 21(8):2852. https://doi.org/10.3390/s21082852
Taherkhani A, Cosma G, Alani AA et al (2018) Activity recognition from multi-modal sensor data using a deep convolutional neural network [C]. In: Science and information conference. Springer, Cham, pp 203–218. https://doi.org/10.1007/978-3-030-01177-2_15
Twomey N, Diethe T, Kull M et al (2016) The SPHERE challenge: Activity recognition with multimodal sensor data [J]. arXiv preprint arXiv:1603.00797. https://doi.org/10.48550/arXiv.1603.00797
Woznowski P, Burrows A, Diethe T et al (2017) SPHERE: a sensor platform for healthcare in a residential environment [M]. In: Designing, developing, and facilitating smart cities. Springer, Cham, pp 315–333. https://doi.org/10.1007/978-3-319-44924-1_14
Wu D, Wang Z, Chen Y et al (2016) Mixed-kernel based weighted extreme learning machine for inertial sensor based human activity recognition with imbalanced dataset [J]. Neurocomputing 190:35–49. https://doi.org/10.1016/j.neucom.2015.11.095
Xie W, Liang G, Dong Z et al (2019) An improved oversampling algorithm based on the samples’ selection strategy for classifying imbalanced data [J]. Math Probl Eng 2019. https://doi.org/10.1155/2019/3526539
Zhao J, Jin J, Chen S et al (2020) A weighted hybrid ensemble method for classifying imbalanced data [J]. Knowl-Based Syst 203:106087. https://doi.org/10.1016/j.knosys.2020.106087
Acknowledgments
This work was supported by the National Natural Science Foundation of China under Grant 61772248. The project name is Sensing Theory and Key Technologies of Big Behavioral Data for Wearable Computing. (Corresponding author: Zhan Huan).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that there are no competing interests regarding the publication of this paper.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Liu, Y., Li, Z., Huan, Z. et al. Daily unbalanced action recognition based on active learning. Multimed Tools Appl 83, 16255–16274 (2024). https://doi.org/10.1007/s11042-023-16181-4
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11042-023-16181-4