An Early Stage Determination of Colon Cancer Through Deep Neural Network

Identification and prediction of Colon cancer have always been difficult and essential tasks for medical professionals. Hospitals are providing costly therapies and operations to treat the illness. So, diagnosing colon cancer at an early stage will be helpful to patients around the world. Currently, the recent improvements in medical technologies, Deep Learning techniques exhibit a major role in predicting the development of colon cancer. In this research work, the risk factors that produce colon cancer-causing genes are selected using feature selection methods such as Fisher Score Univariate Filter (FSUF) and ReliefF Multiclass Filter (RFMF), which are applied separately and integrate the selected prominent genes to minimize the dimensionality of the data. To predict colon cancer, Deep Neural Network (DNN) classifier is utilized on the reduced data and computes the classifier's performance using evaluation metrics. Finally, the outcomes of experiments are compared with two machine learning algorithms, like Independence Bayes(IB) classification model and the k-nearest neighbors(k-NN) classifier. The experimental outcomes demonstrate that the research model significantly reduces the dimension of data space and produced an accurate prediction of colon cancer when compared with other classifiers.