This study specifically examines three ML classification models, namely Fine Tree, Binary GLM (Generalized Linear Model) Logistic Regression, and Efficient Logistic Regression. These models were trained and tested based on an open-access dataset obtained from Kaggle. The dataset contains multiple risk factors and diagnostic parameters that facilitate predictive modeling of lung cancer severity. To determine the effectiveness of the models, strict validation methods were utilized and evaluation measures were utilized. Of the validated models, the Binary GLM Logistic Regression model achieved the highest validation accuracy of 92.9%. In addition, all three models showed effective discrimination ability, as their respective ROC and PR curve evaluations were satisfactory. These results indicate that ML-based predictive models could be advantageous tools in the support of early-stage detection and risk estimation, potentially improving clinical decision making and patient prognosis.