Lung Cancer Detection using Machine Learning: A Comprehensive Approach with Logistic Boosting Model Accuracy

The rise in global cancer rates and the quest for effective treatments are two of the biggest challenges in sustainable healthcare. The role of machine learning in revolutionizing oncology by predicting the optimal treatment plans is crucial. Using a dataset comprising 1000 observations encompassing key medical parameters (Patient Info, Environmental Factors, Genetic and Health Risks, Smoking Status, Symptoms, Severity), four distinct machine learning models were evaluated to forecast suitable treatment plans based on specific patient profiles by employing a voting ensemble strategy. The results showed varying accuracies: the Decision Table led by 95.12%, closely followed by Bayes Net at 97.29%, and Logistic Boosting with an impressive 99.19%. By contrast, the simpler AdaBoost model exhibited a lower accuracy of 80.48%. These findings highlight the superiority of complex ensemble methods, such as decision tables and Bayes nets, in comprehending intricate relationships between genetic, environmental, and clinical factors without sacrificing accuracy.