Advanced Colorectal Cancer Prediction Using CNN and XGBoost Algorithm
A. Kavitha, Gokul Chandrasekaran, S Agilan · 2024
Despite advances in medical prognosis and early cancer detection, colorectal cancer (CRC) remains the third most deadly cancer in the world. Deborah L. Davidson further asserts the need for better models of prediction. This research proposes a hybrid model of CRC prediction that employs a combination of medical image analysis - using convolutional neural network - and structured clinical data processing, which is a tool called XGBoost. It uses CNN model which is trained on colonoscope and histopathological images to detect any cancerous pattern and XGBoost to process patient's demographics, genetic markers and lifestyle factors. A combined feature set is created, which incorporates both the CNN and the XGBoost features, thus allowing the ultimate classifier to evaluate the chances of disease progression.