PCA-based multivariate anomaly detection in mobile healthcare applications

Amor, Lamia Ben; Lahyani, Imene; Jmaïel, Mohamed

doi:10.1109/distra.2017.8167682

Cited by 6 publications

(5 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similarly, these representations can prove to be useful not only in the upper-limb but in other physiological domains. For example, in the past, PCA has been used to evaluate and detect abnormalities in breathing [53] and human gait [54,55] and help identify emergency situations in patients from mobile health application data [56,57]. The overall good and reliable performance of the Leap relative to optical motion capture in participants with and without disabilities in this study suggests that these hand-tracking technologies provide promising and inclusive strategies to deploy in rehabilitation and assistive technology.…”

Section: Looking Aheadmentioning

confidence: 79%

Accuracy of Video-Based Hand Tracking for People With Upper-Body Disabilities

Portnova-Fahreeva,

Yamagami,

Robert-Gonzalez

et al. 2024

IEEE Trans. Neural Syst. Rehabil. Eng.

View full text Add to dashboard Cite

Utilization of hand-tracking cameras, such as Leap, for hand rehabilitation and functional assessments is an innovative approach to providing affordable alternatives for people with disabilities. However, prior to deploying these commercially-available tools, a thorough evaluation of their performance for disabled populations is necessary. In this study, we provide an in-depth analysis of the accuracy of Leap's hand-tracking feature for both individuals with and without upper-body disabilities for common dynamic tasks used in rehabilitation. Leap is compared against motion capture with conventional techniques such as signal correlations, mean absolute errors, and digit segment length estimation. We also propose the use of dimensionality reduction techniques, such as Principal Component Analysis (PCA), to capture the complex, high-dimensional signal spaces of the hand. We found that Leap's handtracking performance did not differ between individuals with and without disabilities, yielding average signal correlations between 0.7-0.9. Both low and high mean absolute errors (between 10-80mm) were observed across participants. Overall, Leap did well with general hand posture tracking, with the largest errors associated with the tracking of the index finger. Leap's hand model was found to be most inaccurate in the proximal digit segment, underestimating digit lengths with errors as high as 18mm. Using PCA to quantify differences between the high-dimensional spaces of Leap and motion capture showed that high correlations between latent space projections were associated with high accuracy in the original signal space. These results point to the potential of low-dimensional representations of complex hand movements to support hand rehabilitation and assessment.

show abstract

Section: Looking Aheadmentioning

confidence: 79%

Accuracy of Video-Based Hand Tracking for People With Upper-Body Disabilities

Portnova-Fahreeva,

Yamagami,

Robert-Gonzalez

et al. 2024

IEEE Trans. Neural Syst. Rehabil. Eng.

View full text Add to dashboard Cite

show abstract

“…Common techniques include principal component analysis [44,45], independent component analysis [46], and autoencoders [47]. These methods can automatically generate new features or reduce data dimensionality while preserving essential information, proving valuable for a wide range of ML and DL models [48][49][50]. For example, [51] applied principal component analysis to data derived from a socioeconomic questionnaire regarding barriers to healthcare.…”

Section: Fementioning

confidence: 99%

Symbolic regression as a feature engineering method for machine and deep learning regression tasks

Shmuel,

Glickman,

Lazebnik

2024

Mach. Learn.: Sci. Technol.

View full text Add to dashboard Cite

In the realm of machine and deep learning regression tasks, the role of effective feature engineering (FE) is pivotal in enhancing model performance. Traditional approaches of FE often rely on domain expertise to manually design features for machine learning models. In the context of deep learning models, the FE is embedded in the neural network's architecture, making it hard for interpretation. In this study, we propose to integrate symbolic regression (SR) as an FE process before a machine learning model to improve its performance. We show, through extensive experimentation on synthetic and 21 real-world datasets, that the incorporation of SR-derived features significantly enhances the predictive capabilities of both machine and deep learning regression models with 34-86% root mean square error (RMSE) improvement in synthetic datasets and 4-11.5\% improvement in real-world datasets. In an additional realistic use case, we show the proposed method improves the machine learning performance in predicting superconducting critical temperatures based on Eliashberg theory by more than 20% in terms of RMSE. These results outline the potential of SR as an FE component in data-driven models, improving them in terms of performance and interpretability.

show abstract

“…It is not only useful to speed up approach execution and reduce storage space required within the context of resource constraints but also help approach achieve better performances by removing redundant features and noisy data if any. In our previous work (Ben Amor, Lahyani, & Jmaiel, ), we have discussed many techniques for dimension reduction purposes. Our study showed the benefits of PCA (Jolliffe, ).…”

Section: Proposed Approachmentioning

confidence: 99%

“…Consequently, X t does not behave in the same way as the multivariate observations used to create the eigenspace model. More details about our multivariate anomaly detection can be found in Ben Amor et al () and Ben Amor, Lahyani, and Jmaiel ().…”

Section: Proposed Approachmentioning

confidence: 99%

AUDIT: AnomaloUs data Detection and Isolation approach for mobile healThcare systems

2019

Self Cite

View full text Add to dashboard Cite

Mobile health care systems highly depend on collected physiological data through medical sensors to provide high‐quality care services. However, inaccurate physiological data from sensors pose a major challenge for health care providers when making decisions, whereas an erroneous decision can affect the user's life. We propose, in this paper, an anomalous data detection and isolation approach for mobile health care systems. Our approach, called AUDIT, detects inaccurate measurements in real time and distinguishes between faults or errors and health events. To do so, we propose reduced time and space complexities algorithms based on dimension reduction within the context of resource constraints. Furthermore, a decision algorithm is proposed while exploring the spatio‐temporal correlation between physiological attributes. First, we describe our approach. Then, we give its implementation details. Finally, to demonstrate the effectiveness of our approach, we show different experiments related to its detection performances and its time and space complexities.

show abstract

PCA-based multivariate anomaly detection in mobile healthcare applications

Cited by 6 publications

References 24 publications

Accuracy of Video-Based Hand Tracking for People With Upper-Body Disabilities

Accuracy of Video-Based Hand Tracking for People With Upper-Body Disabilities

Symbolic regression as a feature engineering method for machine and deep learning regression tasks

AUDIT: AnomaloUs data Detection and Isolation approach for mobile healThcare systems

Contact Info

Product

Resources

About