In this paper, we present the concept and the prototype implementation of our novel “Smart Observatory of Involuntary Medical Seizures (SOIMS)”. SOIMS merges Wireless Body Area Networks (WBAN), Internet of Things (IoT) and Machine Learning (ML) as an intelligent platform for the prediction and modelling of involuntary seizures. The prediction process is elaborated with our proposed algorithms, namely, Qualifying Linear Regression Algorithm (QuLRA), Selective Clustering Algorithm (SeCA) and Real Time Clusters Correlation Algorithm (RT2CA). The assessment of the proposed system is validated based on the Physionet ECG patients’ dataset. The implementation of the prototype involves an IoT/WEB proxy security embedded for translation between nodes CoAP/DTLS protocol and Hospital Information System (HIS) HTTP/TLS protocol. Our proposed solution outperforms existing schemes in the literature at different levels, namely: a) it uses a hierarchical combination of machine learning and prediction algorithms; b) it is open-source, interoperable and user friendly; c) it is a secured prototype implementation; and d) it reaches a higher rate of accuracy according to the correlation criterion.

在本文中，我们介绍了我们的新型“非自愿性医学癫痫发作智能天文台（SOIMS）”的概念和原型实现。SOIMS将无线人体局域网（WBAN），物联网（IoT）和机器学习（ML）合并为智能平台，用于非自愿性癫痫发作的预测和建模。利用我们提出的算法，即合格线性回归算法（QuLRA），选择性聚类算法（SeCA）和实时聚类相关算法（RT2CA），详细说明了预测过程。基于Physionet ECG患者的数据集验证了所提出系统的评估。原型的实现涉及为节点CoAP / DTLS协议与医院信息系统（HIS）HTTP / TLS协议之间的转换而嵌入的IoT / WEB代理安全性。我们提出的解决方案在不同层面上优于文献中的现有方案，即：a）它使用了机器学习和预测算法的分层组合；b）它是开源的，可互操作的且用户友好的；c）这是一个安全的原型实现；d）根据相关准则达到较高的准确率。