Towards Semantic Data Integration in Resource-Limited Settings for Decision Support on Gait-Related Diseases

Daramola, Olawande; Moser, Thomas

doi:10.4018/978-1-7998-6697-8.ch012

Cited by 2 publications

(1 citation statement)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The predictions help the physicians to advice the patients according to the health risks identi ed through intelligent models. The AiEngine in the AiDHealth architecture is dedicated for providing arti cial intelligence for the healthcare data analytics [14].…”

Section: Introductionmentioning

confidence: 99%

AiDHealth: An AI-enabled Digital Health Framework for Connected Health and Personal Health Monitoring

Memon

Abbasi

Rahu

et al. 2023

Preprint

View full text Add to dashboard Cite

We live in a digitally connected world inspired by state-of-the-art ICT technologies and networks, inasmuch as the use of digital gadgets and apps is exponentially increasing in all domains of life. In parallel, artificial intelligence has evolved as an essential tool in all sorts of applications and systems such as healthcare systems. Healthcare is the key domain where the use of ICT infrastructure, technologies and artificial intelligence are playing a major role in providing connected and personalized digital health experiences. The vision is to provide intelligent and customized digital health solutions and involve the masses in personal health monitoring. This research proposes AiDHealth as an intelligent personal health monitoring framework based on artificial intelligence for healthcare data analytics and connectivity for personal health monitoring. AiDHealth relies on various machine learning and deep learning models for achieving prediction accuracy in healthcare data analytics. The extensive Pima Indian Diabetes (PID) dataset has been used for investigation. The findings of our experiments illustrate the effectiveness and suitability of the suggested MLPD model. AdaBoost classifier performance has the highest accuracy in prediction when calculated to the individual classifiers. The AdaBoost classifier produced the best accuracy i.e., 0.975%. The results reveal improvements to state-of-the-art procedures in the proposed model. Next, we trained the models and produced a 10-fold cross-validation illness risk index for each sample. Our findings suggest a need for greater experiments to compare the above-mentioned machine learning methods. We identified the AdaBoost classifier and Decision Tree classifiers with the best prediction with an average of 0.975% and a work Curve Area (AUC) of 0.994%. Thus, because the design of the AdaBoost classifier is superior, it can forecast the danger of type 2 diabetes more accurately than the existing algorithms and lead to intelligent prevention and control of diabetes.

show abstract

Section: Introductionmentioning

confidence: 99%