Lung Cancer Prediction and Risk Assessment: A Machine Learning Approach Integrating Symptoms and Etiological Factors

Lung cancer requires accurate risk assessment and early detection due to its high prevalence and fatality rate. The simultaneous analysis of clinical symptoms and etiological factors using Machine Learning revolutionizes our understanding of lung cancer and patient outcomes. Machine Learning models are built on comprehensive patient data, including lifestyle, symptoms, and 12 variables. We used ten models for precise lung cancer diagnosis, including Decision Tree (DT), K-nearest neighbors (KNN), Support Vector Machine (SVM), Random Forest (RF), Logistic Regression (LR), Naive Bayes (NB), XGBoost, Gradient Boost, CatBoost and AdaBoost. At the National Institute of Cancer Research & Hospital (NICRH) in Mohakhali, TB Gate Road, Dhaka, Bangladesh, we collected 346 data directly from patients using questionnaires, with permission from the director. For every model, we performed parameter tuning to determine the ideal ratio of complexity to accuracy. Notably, with the best accuracy of 98.43% and an excellent area under the curve (AUC) of 0.983, the Decision Tree (DT) performed better than any other classifier. The motivation behind this research is to predict the risk assignment of lung cancer using various machine learning models.