Advancements in Early Detection of Lung Cancer in Public Health: A Comprehensive Study Utilizing Machine Learning Algorithms and Predictive Models

Abstract

Lung cancer stands as the leading cause of death in the United States, attributed to factors such as the spontaneous growth of malignant tumors in the lungs that can metastasize to other parts of the body, posing severe threats. Notably, smoking emerges as a predominant external factor contributing to lung problems and ultimately leading to lung cancer. Nevertheless, early detection presents a pivotal strategy for preventing this lethal disease. Leveraging machine learning, we aspire to develop robust algorithms capable of predicting lung cancer at its nascent stage. Such a model could prove instrumental in aiding physicians in making informed decisions during the diagnostic process, determining whether a patient necessitates an intensive or standard level of diagnosis. This approach holds the potential to significantly reduce treatment costs, as physicians can tailor the treatment plan based on accurate predictions, thereby avoiding unnecessary and costly interventions. Our goal is to establish a sustainable model that accurately predicts the disease, and our findings reveal that XGBoost outperformed other models, achieving an impressive accuracy level of 96.92%. In comparison, LightGBM, AdaBoost, Logistic Regression, and Support Vector Machine achieved accuracies of 93.50%, 92.32%, 67.41%, and 88.02%, respectively.