EnLung: An Ensemble Learning-Based Approach for Enhancing Lung Cancer Prediction

Predicting the early symptoms of several diseases, including lung cancer, is crucial for facilitating rapid diagnosis, treatment, and medication, as the disease is a major cause of mortality globally. Machine Learning (ML) significantly contributes to healthcare by enabling physicians to identify individuals at risk of disease progression, hence facilitating early diagnosis and appropriate therapy interventions tailored to patients. This work presents an ensemble learning-based methodology, EnLung, aimed at improving lung cancer prediction for enhanced disease detection through a three-stage experimental investigation of various diseases. Initially, we evaluated eleven standard ML methodologies. In the second step, we have incorporated feature selection methodologies, including recursive feature elimination and sequential feature selection, with these twelve traditional ML classifiers. In the third stage, we included ensemble methods, such as soft and hard voting and weighted averaging, to improve the model's effectiveness. The investigations concluded that the increased accuracy, precision, sensitivity, specificity, f-measure, and AUC for lung cancer are <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$99.19 \%, 99.43 \%, 99.43 \%, 98.61 \%, 99.43 \%$</tex>, and 0.976, respectively. The suggested ensemble hybrid ML techniques utilize many predictive performance metrics to demonstrate the efficacy and productivity of the study in a disease diagnosis.