Lung Cancer Prediction Using an Enhanced Neutrosophic Set Combined with a Machine Learning Approach

Abstract

 Lung cancer (LC) remains one of the most lethal diseases globally, necessitating the development
 of advanced predictive models for early detection and accurate diagnosis. Traditional classification techniques
 often struggle with uncertainty and indeterminacy in medical data, which can lead to misdiagnosis and reduced
 diagnostic reliability. To address this issue, we propose an Enhanced Neutrosophic Set (ENS) framework
 integrated with machine learning algorithms to improve the prediction accuracy of lung cancer. Neutrosophic
 Set (NS) theory extends classical and fuzzy logic by introducing three independent membership components:
 truth, indeterminacy, and falsity, which enable more effective modeling of uncertainty in clinical datasets.
 The proposed ENS model enhances decision-making by optimizing feature selection and minimizing ambiguity
 in patient data representation. We apply machine learning classifiers including Logistic Regression (LR), K
Nearest Neighbors (KNN), and Random Forest (RF) to evaluate the performance of the ENS-transformed
 dataset in predicting lung cancer risk. Experimental results indicate that the ENS-based models outperform
 traditional approaches in terms of classification accuracy, sensitivity, and specificity. This study demonstrates
 the effectiveness of neutrosophic-based AI frameworks in medical diagnostics and highlights their potential in
 developing reliable, early detection systems for lung cancer and other critical diseases

DOI: 10.5281/zenodo.15878877