Parallel extreme gradient boosting classifier for lung cancer detection

Abstract

Most lung cancers do not cause symptoms until the disease is in its later stage. That led the lung cancer having a high fatality rate compared to other cancer types. Many scientists try to use artificial intelligence algorithms to produce accurate lung cancer detection. This paper used extreme gradient boosting (XGBoost) models as a base model for its effectiveness. It enhanced lung cancer detection performance by suggesting three stages model; feature stage, XGBooste parallel stage and selection stage. This study used two types of gene expression datasets; RNA-sequence and microarray profiles. The results presented the effectiveness of the proposed model, especially in dealing with imbalanced datasets, by having 100% each of sensitivity, specificity, precision, F1_score, area under curve (AUC), and accuracy metrics when it applied on all of the datasets used in this study.