Proposing an Early Diagnostic Deep Learning Approach to Detect Lung Cancer from Short-Breaths

This paper proposes a deep learning approach that might help to diagnose lung cancer at an early stage. A deep neural network (DNN) will be trained via classification and clustering to predict disparity in the dimensions of the diaphragm. The model will use a rain forest algorithm (RFA) for the initial classification of the clippings. A deep clustering that uses a feed-forward attribute will be implemented for the second-half of the hidden layers. This model will be able to identify short breaths thereby resulting in the early diagnosis of lung cancer. The change in the breathing habits of individual will be highlighted by the trained model further prompting the individual to take remedial actions at a much early phase. The strategy behind the model is creating a scope in the form of an alert via an application or device with the integration of IoT platforms that can be later developed into a business model.