Lung-Cancer stage prediction using machine learning approach

Abstract

Physical illnesses like lung cancer have become more common nowadays. The world nowadays is aware of the subject. The disease known as lung cancer affects most individuals. A measure of the sickness is the variations in diagnostic report ratios between patients who are normal and those who are afflicted. Numerous investigations have already been conducted on the illness of lung cancer. I've identified a few excellent chances to improve the procedure even further. I suggest using effective algorithm models to anticipate hazards and raise early alert. My suggested approach is easy to utilize in practice and appropriate for basic projections of lung cancer sickness. The GitHub website served as the dataset's host. Several five distinct kinds of algorithms, including Decision Trees, KNN, ANN, SVM, and LR (logistic regression), have been used. In order to defend the performances, I additionally used a few other ensemble models. ANN had the accuracy at 99.50%, KNN at 99.50%, and SVM at around 97.50%. After that, I received a perfect score of 100% in both Decision Tree and Logistic Regression. I optimized each classifier's parameters using hyper-parameter tweaking. The experimental research analyzed the findings of other recent studies and produced more accurate estimates of lung cancer disease, with 100% accuracy being the greatest performance.