Lung cancer detection using machine learning methods

Abstract

Cancer-related death is now more common than ever with cancer of the lungs being the leading cause of it. In such regards, the survival percentage for individuals with lung cancer must be increased through early identification. In this paper, we propose a machine learning-based approach for lung cancer detection using data from patients. The proposed method has three stages: preprocessing, feature extraction, and classification. In preprocessing, we extracted the necessary data from the dataset we collected. In the feature extraction stage, we processed the raw data while keeping the original information unchanged, in order for the classifier to work on. Finally, in the classification stage, we use 6 different classifier models to classify the extracted features. We evaluated the proposed method on a publicly available dataset from Kaggle of lung cancer causes and age. The results of the trial confirm the feasibility of the suggested strategy to achieve high accuracy, sensitivity, specificity, and AUC score. In conclusion, using machine learning, The suggested approach offers a powerful and efficient way to detect lung cancer. The proposed method can potentially assist in the detection of lung cancer, thereby improving the survival rate of lung cancer patients.