Comparative Analysis of Machine Learning Models for Lung Cancer Prediction Using Radiographic Images

Abstract

There has been an upsurge in objective to promote and enhancing lung cancer screening techniques as people become more aware of the behaviours and risks associated with lung cancer. Machine learning-based lung cancer prognosis models were proposed to assist doctors for incidental or screen found ambiguous pulmonary nodules. Using these technologies, the variability in tumour classification might be reduced and decision-making could be improved, resulting in fewer benign nodules being followed up on. For the main purpose of lung cancer prediction, we conduct comparative research of important machine learning algorithms in this paper and present statistical proofs that certain algorithms will perform better for radiographic-based detection than others for this purpose. In conjunction with machine learning approaches, pre-trained architecture like VGG provides a good detection for biomarkers, making it a viable tool for illness classification. They have the ability to classify while also reducing the number of false positives accurately.