Chapter 27 - Comparison of machine learning models for lung cancer prediction using different feature selection methodologies

Abstract

Lung cancer is one of the most widespread diseases with significant fatality rates worldwide. Machine learning (ML) algorithms have recently demonstrated great promise for predicting lung cancer. The proposed research focuses on the extraction of helpful features from patient data that can enhance the precision and understandability of machine learning models. Correlation-based feature selection, recursive feature elimination (RFE), and tree-based feature selection are examined to see which is the most effective feature selection technique for predicting lung cancer. The ML models Naive Bayes (NB), K-nearest neighbor (KNN), and support vector machines (SVMs) were trained to produce predictions using the chosen features. Accuracy, precision, recall, and F1-score metrics were used to assess the model's performance. When the models were trained with and without feature selection, KNN and SVM displayed the highest prediction accuracy compared with NB. One of the feature selection techniques that has helped machine learning models to be trained with the most relevant attributes, increasing prediction accuracy, is tree-based feature selection.