Machine Learning Models in Bone Tumor Diagnosis: Performance Evaluation and Hyperparameter Optimization

Bone tumor is a very severe condition, it can be benign or malignant. Provides major difficulties in medical diagnosis. For diverse kinds of bone cancers, conventional diagnostic techniques including physical exams and plain radiography often fall short. Although useful, advanced imaging methods such as MRIs and CT scans still have restrictions because of interpretation problems and professional variances. Then correct treatment and better patient outcomes can be helpful in early and precise diagnosis, by using distinct Machine Learning approaches diagnosis might be improved by early prediction of bone tumors. This work investigates distinct methods of Machine Learning, such as XGBoost, K-Nearest Neighbors (KNN), and Random Forest (RF) classifiers. To improve model performance, the work also optimizes the hyperparameters using grid search. The results showed that hyperparameter adjustment greatly raised accuracy. Grid search gave an accuracy of $85 \%$, this study underlines how Artificial Intelligence and Machine Learning could solve the shortcomings of conventional diagnosis techniques and raise the precision and efficiency of bone tumor identification.