Since the end of 2019, the world has faced severe issues over Corona Virus Disease of 2019 (COVID-19). So there is a need for some essential precautionary measures until the development of vaccines to battle the COVID-19 pandemic. In addition to that, quarantine and social distancing have become the more significant practice in the world. COVID-19 associated with the virus not only degraded the economy of the world due to the lockdown but also saturated the healthcare system of the people due to its exponential spread. In this case, the Internet of Things (IoT) system offers frequent monitoring facilities to doctors. But, the fight against COVID-19 gets continued until people get vaccinated. Therefore, an IoT-based COVID prediction is designed using deep learning techniques. Firstly, IoT data is collected from online resources. Then, the data is fed to the autoencoder (AE) for attaining the deep features. Further, the deep features are forwarded to the attentive and adaptive ensemble model (AAEM), which includes deep temporal convolution network (DTCN), one-dimensional convolutional neural network (1DCNN), and long short-term memory (LSTM) model for COVID prediction or monitoring. By utilizing the hybrid algorithm fitness position of Eurasian oystercatcher and sewing training (FPEOST), the parameter in the ensemble model is tuned for further improvement in the process. Finally, the COVID-19 disease prediction outcomes are attained on the basis of the high-ranking process. Thus, the developed model achieved an effective prediction rate than conventional approaches over multiple experimental analyses.

自 2019 年底以来，世界面临 2019 年冠状病毒病（COVID-19）的严重问题。因此，在研制出对抗 COVID-19 大流行的疫苗之前，需要采取一些必要的预防措施。除此之外，隔离和社交距离已成为世界上更重要的做法。与该病毒相关的 COVID-19 不仅因封锁而导致世界经济恶化，而且由于其指数级传播而使人们的医疗保健系统饱和。在这种情况下，物联网（IoT）系统为医生提供频繁的监控设施。但是，抗击 COVID-19 的斗争将继续下去，直到人们接种疫苗为止。因此，基于物联网的新冠肺炎预测是利用深度学习技术设计的。首先，物联网数据是从在线资源收集的。然后，数据被输入到自动编码器（AE）以获得深层特征。此外，深度特征被转发到注意力和自适应集成模型（AAEM），其中包括深度时间卷积网络（DTCN）、一维卷积神经网络（1DCNN）和针对新冠病毒的长短期记忆（LSTM）模型预测或监测。通过利用欧亚蛎鹬适应位置与缝纫训练的混合算法（FPEOST），对集成模型中的参数进行调整，以进一步改进该过程。最后，在高排序过程的基础上获得了COVID-19疾病预测结果。因此，在多次实验分析中，所开发的模型比传统方法实现了有效的预测率。