<![CDATA[Purpose of the study: This study aims to evaluate and compare the performance of four machine learning algorithms Naïve Bayes, Logistic Regression, Decision Tree, and Random Forest for bone tumor grade classification using structured clinical data and to identify the most effective algorithm for supporting diagnostic decision-making in orthopedic oncology. Methodology: An experimental quantitative research design was employed using a publicly available Bone Tumor dataset from Kaggle containing 500 records. Model development was conducted in Google Colaboratory using Python and Scikit-learn. The evaluated algorithms included Naïve Bayes, Logistic Regression, Decision Tree, and Random Forest. Data preprocessing, feature selection, and train-test splitting (80:20) were performed. Model performance was assessed using accuracy, precision, recall, and F1-score metrics. Main Findings: The results demonstrated that all machine learning models were capable of classifying bone tumor grades with satisfactory performance. Logistic Regression achieved the best overall performance, obtaining 81% accuracy, precision values of 79–82%, recall values of 73–86%, and F1-scores of 76–84%. Decision Tree and Naïve Bayes showed moderate performance, while Random Forest exhibited reduced testing performance despite strong training results, indicating overfitting and lower generalization capability. Novelty/Originality of this study: This study contributes a comprehensive comparison of classical machine learning algorithms for bone tumor grade classification using structured clinicopathological data rather than imaging data. The findings demonstrate that interpretable models such as Logistic Regression can achieve reliable predictive performance, providing an accessible and computationally efficient alternative for clinical decision-support systems in resource-limited healthcare.]]>
Copyrights © 2026
