Many researchers have recently considered patients’ health and provided an optimal and appropriate solution. With the advent of technologies such as cloud computing, Internet of Things and 5G, information can be exchanged faster and more securely. The Internet of things (IoT) offers many opportunities in the field of e-health. This technology can improve health services and lead to various innovations in this regard. Using cloud computing and IoT in this process can significantly improve the monitoring of patients. Therefore, it is important to provide a useful method in the medical industry and computer science to monitor the status of patients using connected sensors. Thus, due to its optimal efficiency, speed, and accuracy of data processing and classification, the use of cloud computing to process the data collected from remote patient sensors and IoT platform has been suggested. In this paper, a prioritization system is used to prioritize sensitive information in IoT, and in cloud computing, LSTM deep neural network is applied to classify and monitor patients’ condition remotely, which can be considered as an important innovative aspect of this paper. Sensor data in the IoT platform is sent to the cloud with the help of the 5th generation Internet. The core of cloud computing uses the LSTM (long short-term memory) deep neural network algorithm. By simulating the proposed method and comparing the obtained results with other methods, it is observed that the accuracy of the proposed method is 97.13%, which has been improved by 10.41% in average over the other methods.

Patients' health and providing an optimal and appropriate solution have recently been considered by many researchers. With the advent of technologies such as cloud computing and 5G network, information can be exchanged faster and more securely. Using cloud computing in this process can significantly improve the monitoring of certain patients. Therefore, providing a favorable method in the medical industry and computer science to monitor the status of patients using connected sensors is very important. Thus, due to its optimal efficiency, speed, and accuracy of data processing and classification, the use of cloud computing to process the data collected from remote patient sensors and the Internet of Things (IoT) platform has been suggested. In this paper, a prioritization system is used to prioritize sensitive information in IoT, and in cloud computing, LSTM deep neural network is used to classify and monitor patients' condition remotely, which can be considered as an important innovative aspect of this paper. Sensor data in the IoT platform is sent to the cloud with the help of the 5th generation Internet. The core of cloud computing uses the LSTM deep neural network algorithm. Through simulating the proposed method and comparing the obtained results with other methods, it was observed that the accuracy of the proposed method has been improved significantly compared to other methods.

High amounts of patient and healthcare data are generated daily on the Internet of Things (IoT). The processing time and analyzing the big data received from IoT devices, as well as providing the necessary accuracy for data classification, are considered as the most important challenges in IoT. Therefore, this study aims to improve the accuracy of data classification by implementing hybrid approaches of feature selection (FS) and deep learning (DL) within the fog computing infrastructure while reducing the response time (RT) and bandwidth usage. Also, the major goal is to analyze and classify the states of patients remotely through IoT, cloud computing and fog computing technologies. The information processing time is reduced by implementing Particle Swarm Optimization (PSO) and Imperialist Competitive Algorithm (ICA) in fog computing to select the prominent features, and the classification accuracy is improved through classifying new data by a deep neural network (DNN) model. The simulation results show that the accuracy of the classification and remote monitoring of patients is 98.54%, which is about 4.5% better than the case without using PSO-ICA algorithms, and it is also improved by about 10% on average compared to other methods such as Linear Regression (LR), K Nearest Neighbors (KNN), Neural Network (NN), and Bayesian Belief Net (BBN).