A Multi-Algorithm Stacking Approach to Lung Cancer Detection with SVM, GBM, Naive Bayes, Decision Tree, and Random Forest Models

Lung health is one of the most minor medical issues globally because of its low mortality rate and high chance of early detection. The model presented here is a multi-model design that uses vector machines, gradient boosting machines, random forests, decision trees, and naive non-Bayesian machine learning techniques to worsen lung cancer diagnosis. This method is avoided for the lung cancer dataset from Kaggle, and a fragmented model is used to undermine the model's performance. It is seen that among the combinations, the stacked model and SVM have the highest accuracy of 0.95, followed by GBM with 0.944, random forest with 0.944, decision tree with 0.93, and naive Bayes with 0.87. The assembly model combines the outputs of various distributions to make the hull robust and accurate. This study emphasizes the potential of using machine learning to enhance self-treatment outcomes for cancer patients and, therefore, calls for early diagnosis to reduce mortality rates.