Detection of Lung and Colon Cancer using Average and Weighted Average Ensemble Models

Abstract

INTRODUCTION: Cancer is a life-threatening condition triggered by metabolic irregularities or the convergence of hereditary disorders. Cancerous cells in lung and colon leads more death rate count in the human race today. The histological diagnosis of malignant cancers is critical in establishing the most appropriate treatment for patients. Detecting cancer in its early stages, before it has a chance to advance within the body, greatly reduces the risk of death in both cases.
 OBJECTIVES: In order to examine a larger patient group more efficiently and quickly, researchers can utilize different methods of machine learning approach and different models of deep learning used to speed up the detection of cancer.
 METHODS: In this work, we provide a new ensemble transfer learning model for the rapid detection of lung and colon cancer. By ingtegrating various models of transfer learning approach and combining these methods in an ensemble, we aim to enhance the overall performance of the diagnosis process.
 RESULTS: The outcomes of this research indicate that our suggested approach performs better than current models, making it a valuable tool for clinics to support medical personnel in more efficiently detecting lung and colon cancer.
 CONCLUSION: The average ensemble is able to reach an accuracy of 98.66%, while the weighted-average ensemble with an accuracy of 99.80%, which is good with analysis of existing approaches.