Ensemble Machine Learning Techniques for Pancreatic Cancer Detection
H V Ramachandra; Pundalik Chavan; Anooja Ali; H C Ramaprasad
Pancreatic cancer is a disease with high mortality rates that closely correlate with its incidence. Patients with pancreatic cancer often do not experience symptoms until the disease has progressed to an advanced stage. There is no established screening program for individuals at high risk of developing pancreatic cancer. Research conducted earlier has discovered that a combination of three protein biomarkers (LYVE1, REG1A, and TFF1) present in urine can aid in the identification of significant Pancreatic Ductal Adeno Carcinoma (PDAC). The RIC-GD method is a novel machine learning approach proposed for the detection of pancreatic tumors. It utilizes an ensemble classifier to enhance the classification performance. The technique involves using a set of classifiers and determining the similarity measure between the training and testing samples to ensure accurate classification of the samples. The accuracy and specificity of the RIC-GD method have been evaluated and compared to Naive Bayes and decision tree methods. The results demonstrate that the RIC-GD method achieves an accuracy of 92% than the other methods.