Lung Cancer Detection Using Multiple Machine Learning Algorithms

Lung cancer remains one of the leading causes of cancer-related mortality worldwide, primarily due to late-stage diagnosis. Early and accurate detection through medical imaging can significantly improve patient outcomes. This project proposes an advanced ensemble learning approach for lung cancer detection using chest X-ray and CT scan images. Two publicly available Kaggle datasets were combined and preprocessed to classify images into two categories, cancer or no cancer, with a 70:30 train-test split. Five high-performing convolutional neural network (CNN) models—ResNet-50, DenseNet, NASNet, InceptionV3Large, and Xception—were fine-tuned by adding two custom layers for binary classification. Each model was trained independently and used to generate binary predictions, which were then combined using a majority voting mechanism. This ensemble strategy effectively utilised the complementary strengths of individual models, resulting in higher accuracy and robustness than any single model. The study demonstrates the potential of ensemble deep learning techniques in building scalable and reliable tools for early lung cancer diagnosis.