Predictive Analytics in Oncology: A Comprehensive Study on Lung Cancer Risk Factors and Machine Learning Model Performance

Cancer still remains one of the most fatal diseases in this world. Early detection and diagnoses can easily enhance treatment outcomes and survival probability by a great level. This paper presents the leverage of advanced machine learning techniques for cancer risk prediction using which timely intervention can be enabled. A comprehensive cancer dataset of 11 features capturing demographics, lifestyle habits, family history etc. was used to train prediction models based on logistic regression, random forests and neural networks. Extensive evaluations demonstrate the superiority of the neural network with near perfect precision, recall and F1-scores. The sensitivity analysis brought some of the most influential risk factors forward. The predictive performance proved the potential of machine learning in preventive oncology with data-driven screening recommendations and personalized care. Future work should focus on integrating di-verse health data and testing model performance in real clinical research.