Performance Comparison of Ensemble Learning and Deep Learning Models for Lung Cancer Detection in CT Imaging
A. Anto Sagaya Priscilla, R. Balamanigandan
Abstract
Lung cancer detection and classification tasks on CT scans are achieved by implementing ensemble learning methods with convolutional neural networks (CNNs). CT image modification consists of preprocessing steps followed by filtering protocols and segmentation techniques which together create improved feature extraction for precise classification methods. A three-layered ensemble learning system consisting of decision tree, random forest, and gradient boosting methods detects concept behavior, achieving a 89% accuracy rate. The ensemble model shows superior robustness and precision for lung cancer discrimination and reaches a 92% accuracy with reduced overfitting and noise sensitivity. This indicates that ensemble learning is better than the CNN for reliable lung cancer detection in medical images.