A Novel CNN-TLSTM Approach for Dengue Disease Identification and Prevention using IoT-Fog Cloud Architecture

Abstract

One of the mosquito-borne pandemic viral infections is Dengue which is mostly transmitted to humans by the Aedes agypti or female Aedes albopictis mosquitoes. The dengue disease expansion is mainly due to the different factors such as climate change, socioeconomic factors, viral evolution, globalization, etc. The unavailability of certain antiviral therapy and specific vaccine increases the risk of the dengue disease spreading even further. This arises the need for a novel technique that overcomes the complexities associated with dengue disease prediction such as low reporting level, misclassification, and incompatible disease monitoring framework. This paper mainly overcomes the above-mentioned problems by integrating the Internet of Things (IoT), fog-cloud, and deep learning techniques for efficient dengue monitoring. A compatible disease monitoring framework is formed via the IoT devices and the reports are effectively created and transferred to the healthcare facilities via the fog-cloud model. The misdiagnosis error is overcome in this paper using the novel Hybrid Convolutional Neural Network (CNN) with Tan h Long and Short Term Memory (TLSTM) based Adaptive Teaching Learning Based Optimization (ATLBO) algorithm. The ATLBO optimized CNN-TLSTM architecture mainly analyzes the dengue-related parameters such as Soft Bleeding, Muscle Pain, Joint Pain, Skin rash, Fever, Water Site, Carbon Dioxide, Water Site Humidity, Water Site Temperature, etc. for an efficient clinical decision making and timely disease diagnosis. The experimental results are conducted using a real-time dataset and its performance is validated using various performance metrics. When compared in terms of different statistical parameters such as accuracy, f-measure, mean square error, and reliability, the proposed method offers superior results in the case of dengue disease detection than other existing methods. The ATLBO optimized hybrid CNN-TLSTM shows an accuracy of 96.9%, a precision of 95.7%, recall of 96.8%, and F-measure of 96.2% which is relatively high when compared to the existing techniques. The proposed model is capable of identifying the patients in a certain geographical region and preventing the disease emergency via immediate disease diagnosis and alerting the healthcare officials to offer the stipulated services.

Mathematical Framework for Wearable Devices in the Internet of Things Using Deep Learning

“…Additionally, wearable IoT devices classify the types of viruses based on blood samples and send the information to local emergency rooms. So, to figure out how far the emergency rooms and hospitals are from each other [ 5 ], we can use Equation (1). …”

“…In Equation (1), the objective function where the identification distance is maximized. Equation (2) can be used to define the energy representation of local computation tasks when wearable devices with unloading tasks are used [ 5 , 21 ]. …”

“…Equation (7) is used for wearable devices to figure out the weighted combination of variables or quality factors since different parameters are used in the representation process [ 5 ]. …”

“…Even though the working model has been fully sensitized, a significant flaw is that not all applications have access to the fabrication conditions in a proper manner. An offloading strategy that boosts the energy of node systems has been developed to address some of the difficulties in fabrication processes [ 5 ]. Smart wearable gadgets have a greater capacity for storing significant data amounts and have strong offloading techniques.…”

DengueFog: A Fog Computing-Enabled Weighted Random Forest-Based Smart Health Monitoring System for Automatic Dengue Prediction

“…Between centralized cloud infrastructure and the IoT is a stratum of fog computing for managing messaging overhead, latency, decision-making, resident storage, and information preprocessing problems between the end user and the cloud server [ 35 , 36 , 37 ]. Integration of IoT, fog computation, and the cloud enhances scalability and agility for controlling the mosquito-borne diseases by incorporating topographical areas and assessments based on real-world data analytics [ 38 ].…”