DeepELR: Deep learning-based energy and link stability aware routing in IoT for heart disease classification

Vidya C.A., V.Baby Shalini,
DeepELR: Deep learning-based energy and link stability aware routing in IoT for heart disease classification,
Computational Biology and Chemistry,
Volume 119,
2025,
108574,
ISSN 1476-9271,
https://doi.org/10.1016/j.compbiolchem.2025.108574.
(https://www.sciencedirect.com/science/article/pii/S147692712500235X)
Abstract: Internet of Things (IoT) is an advancing trend that is likely to dominate future-generation technology. Predominant aim of the research is that objects and devices can be connected uniquely identifiable smart objects. IoT can involve all of those objects that are so profoundly integrated into the environment. Among the many applications, IoT health monitoring is one of the most popular in wearable electronics, making intelligent healthcare monitoring highly beneficial to heart disease patients. While the most important aspect of IoT is its integration with intelligent healthcare monitoring, it advances remote intelligent computing with an IoT scenario of health tracking. Early diagnosis and predictive modelling based on Machine Learning (ML) have been stated to be highly effective in medical analysis. This research aims at developing heart disease detection and integration into the IoT paradigm with the design of an effective routing algorithm for using data sensed by IoT nodes by medical professionals. A novel routing algorithm, called Deep Learning-based Energy and Link Stability Aware Routing (DeepELR) is developed for successful transmission of patient data to the destination. In this case, a Deep Recurrent Neural Network (DRNN) is used to forecast the nodal energy and link stability ensuring that the routing algorithm does not lose any data.With the transferred data, heart disease classification is done through ASSA-based ensemble learning. Besides, disease classification is performed with Accelerated Sparrow Search Algorithm-based ensemble learning (ASSA-based ensemble learning) technique, where the ensemble methods comprising DRN, DEB and RNN are trained by executing ASSA, which is generated by improving Sparrow Search Algorithm (SSA) by accelerating the algorithm.Hence, from the experiments, it was revealed that the proposed approach attained an energy consumption of 0.762 J with a Packet Delivery Ratio (PDR) of 94.3 % along with link stability of 31.8 %.
Keywords: IoT; Routing; Heart disease; Optimization algorithm; Ensemble technique; Deep learning