A Comparative Evaluation of Logistic Regression, Random Forest, K-Nearest Neighbors, and Support Vector Machine Algorithms for Machine Learning-Based Lung Cancer Risk Classification Using Accessible Clinical, Demographic, and Lifestyle Risk Factors

Abstract

Lung cancer remains one of the leading causes of cancer-related mortality worldwide, largely due to late-stage diagnosis and limited access to effective early screening tools. Although low-dose computed tomography (LDCT) has demonstrated potential for early detection, its high cost and limited availability restrict widespread implementation, particularly in resource-constrained settings. Consequently, there is growing interest in computational approaches that utilize clinical, demographic, and biomarker data to support early lung cancer risk identification. This study investigates the effectiveness of machine learning algorithms in predicting lung cancer risk using accessible patient data and blood-related indicators. A machine learning–based classification framework was developed using the publicly available lung_cancer.csv dataset obtained from Kaggle. The dataset contains demographic information, lifestyle factors, smoking history, environmental exposures, and clinical measurements associated with lung cancer risk. Data preprocessing included feature scaling, normalization using the Min–Max method, and feature selection using the SelectKBest technique to identify the most relevant predictors. The dataset was divided into training and testing sets using a 70:30 ratio. Four classification algorithms—Logistic Regression (LR), Random Forest (RF), k-Nearest Neighbors (KNN), and Support Vector Machine (SVM)—were implemented using Python and the Scikit-learn library. Model performance was evaluated using accuracy, precision, recall, F1-score, and confusion matrix analysis. The experimental results indicate strong predictive performance across all evaluated models, with classification accuracies exceeding 93%. Among the algorithms, the Random Forest model achieved the highest accuracy of 95.0%, along with high precision, recall, and F1-score values. Feature selection identified several influential predictors, including smoking exposure, air pollution index, pulmonary function indicators, and inflammatory biomarkers, highlighting their relevance in lung cancer risk prediction.