Model Compression by EL-XAI in IoMT Systems

IoMT systems are evolving at a higher rate than other IoT applications. As health is of major concern, more health monitoring sensors and devices are incorporated which helps in the well-being of the people at their comfort zone. The IoMT system needs highly capable sensors and devices to know the exact information of the patients from the remote area. On the other hand, usage of high-resolution devices leads to the higher amount of data packets for communication purposes. This increases the workload of the cloud server that can be faced using compression algorithms. This paper proposes the compression technique for various datasets because the medical devices produce different data format outputs. So, it is difficult for the cloud server to understand and interpret the information from the IoMT devices. Therefore, this paper uses an Explainable Artificial Intelligence based Ensemble learning algorithm using Lempel Ziv, 2D and 3D wavelet compression techniques in the cloud server of the IoMT system. The proposed system utilizes the resource efficiently and shown good results in terms of accuracy, throughput, and storage saving. The results are compared with the conventional EL algorithm to study the proposed system performance.