Publication

Abstract : Lung cancer is caused by uncontrolled cell growth in lung tissues, which is determined by smoking, genetic susceptibility, and environmental toxins. Being one of the major causes of cancer death, late-stage diagnosis lowers survival rates, emphasizing the importance of early and accurate risk prediction. This research compares several machine learning models, including feature selection and hyperparameter tuning to maximize performance while addressing class imbalance through synthetic oversampling methods. The assessment proves that an RFXGB classifier yields better predictive accuracy and reliability. Statistical analysis and cross-validation also ascertain the reliability of the proposed framework, reinforcing its value in clinical contexts. The outcomes prove that the ensemble-based method improves accuracy and robustness, making it an effective tool for early detection of lung cancer and hence better patient outcomes.