Hybrid neural network Architectural models for lung Cancer classification

Abstract

Over the past few years, the occurrence of cancer is noted to be newlinevery prominent in the individuals and different types of cancer like blood newlinecancer, cervical cancer, larynx cancer, breast cancer, lung cancer, colon newlinecancer, and prostate cancer and so on are the ones that are likely to occur. newlineCancer detection and classification is the most important task that has to be newlinecarried out at an earlier stage so that it will save numerous human lives. newlineHence, cancer detection and classification has been a prominent research area newlineand researchers work on developing models for early detection of cancer and newlinesubsequently classifying it. newlineNumerous conventional techniques exists that are employed to newlinedetect the lung cancer nodules using image processing techniques. But in newlineorder to be more accurate and perform a better classification with early newlinedetection, the machine learning classifier using its neural network modelling newlineachieves better classification rate. Due to which, the proposed research newlineattempts to model new hybrid neural network architectural models for newlineperforming lung cancer classification and identify the occurrence of possible newlinepulmonary nodules in the lung tissues and thereby can save human lives. In newlinethis work, the developed hybrid models are applied on datasets from lung newlineimage database consortium and that of clinical data samples from hospitals. newlineSimulations are carried out and the metrics are evaluated in respect of newlineclassification process to prove the effectiveness of the developed new newlinemodels. The research contributions made in this thesis are as presented newlinebelow. newline