Performance Analysis of CNN Model for Disease Prediction on Spark and Hadoop

Skin disorders are one of the most dreaded games which humans are equally playing in the whole world and need appropriate and potent diagnostic methods for quick cure. The aim of this work is to implement an intelligent and scalable scheme of skin disease identification that will be based on well-designed Convolutional Neural Networks (CNNs), as such approach will be combined with reliable distributed computing platforms, among which are Hadoop and Spark. The study in which the final part of the report concerning the performance measuring is considered systematically examines the throughput that is obtained at Hadoop and Spark by using the CNN mapper. One key parameter is the throughput value of the system used to demonstrate how fast and accurately the system performs physiological image data processing and arrives at conclusions. The objective of the study is to explore how the CNN mappers ensure enhancement in diagnoses of dermatology and not only that but also the complex dynamics of the Carter throughout the distributed computing environment. This research on the arrangement between CNNs, Spark and Hadoop already open new fields and further start to show cases how machine learning-based technologies totally might become a dominant way of diagnostic systems development in this universe in short times.