Ensembling Naive Bayes and Random Forest for Lung Cancer Prediction

This study enhances the precision of lung cancer predictions by introducing an innovative ensemble approach that merges the capabilities of the Naive Bayes and Random Forest algorithms. The foundation of this method is the integration of Random Forest’s exceptional ability to select relevant features and Naive Bayes’ strength in probabilistic predictions. Initially, Random Forest analyzes and ranks the relevance of features within the lung cancer dataset, identifying the most critical indicators of the disease. The Naive Bayes classifier then utilizes these key features to create a probability matrix, significantly improving lung cancer prediction accuracy. This blend aims to address the limitations of each algorithm when used alone, particularly the assumption of feature independence and the challenge of analyzing complex data sets. The performance of this hybrid technique was verified through comprehensive dataset analysis, demonstrating its effectiveness and superiority as a tool for healthcare prediction. Notably, the combined approach achieved a marked increase in predictive accuracy compared to the individual application of each algorithm. The findings reveal that integrating feature selection with probabilistic modeling substantially enhances the detection accuracy of lung cancer.