Echo state learned compositional pattern neural networks for the early diagnosis of cancer on the internet of medical things platform

Recent years medical fields are facing more challenging issues while detecting disease because day by day disease are developed due to the several factors. The IoT medical device is placed in human body that captures the people’s cancer related health conditions such as breast conditions, skin rashes, teeth related information and lung information is collected. The gathered data analyzed using machine learning, data mining algorithm to get the changes present in the people health condition. Features are determined from the collected information, and optimized features are chosen by the implementation of a selection method for Iterated memetic correlations. The selected features are trained using an echo state deep learning process that uses various levels of hidden layers that solve possible cognitive cancer decisions without making any errors. The new features are classified using the compositional pattern neural network with the assistance of trained features. Finally, system efficiency is assessed using experimental results such as F-measure, mean absolute error, precision, recall and abnormal pattern prediction rate.