Lung Cancer Prediction Using Ensemble Learning

Lung cancer initiates from the uncontrolled growth of abnormal cells within the lungs, presenting a severe health concern that can lead to significant harm and mortality. It stands as the foremost contributor to cancer-related fatalities on a global scale, affecting both men and women with the highest mortality rates. Smoking constitutes the primary factor in roughly 85% of all lung cancer instances. Diagnosis of lung cancer commonly transpires during advanced stages, limiting the available treatment options. Overall, lung cancer remains a predominant cause of cancer-related deaths worldwide, arises from uncontrolled lung cell growth, primarily linked to smoking, which accounts 85% of the cases. Unfortunately, lung cancer often remains undetected until advanced stages, limiting treatment options. In our research investigates a diverse range of methods, encompassing traditional ones like Support Vector Machine(SVM) and Random Forest(Bagging),as well as advanced techniques such as Convolutional Neural Networks (CNN), VGG16 CNN, ResNet50, Boosting (Adaboost and Xgboost), and a unique hybrid approach. Through extensive experimentation and cross-validation, we assess the predictive power, interpretability, and practical viability of these methods in predicting lung cancer, using a rich dataset of clinical attributes and outcomes. Our comprehensive analysis offers insights valuable for both medical practitioners and researchers, contributing to informed decisions for improved lung cancer prediction and highlighting the potential for deep learning and machine learning to revolutionize cytology-based diagnostic procedures in the field of healthcare.