Evaluation of Machine Learning Algorithms for PredictingTumor and Non-tumor Brain Mri Images

Early diagnosis of brain tumors using MRI and artificial intelligence algorithms is fundamental in improving treatment results. MRI images serve as the primary tool for identifying brain tumors. This study aims to evaluate machine learning algorithms for diagnosing brain tumors and non-tumors using MRI images.

From kaggle.com a total of 2400 MRI images were collected, and a pre-processing step was performed on them. Algorithms such as logistic regression, decision tree, random forest, simple Bayes method, support vector machine, and K nearest neighbor were also implemented on the images.

After applying all the algorithms, the values of training accuracy, test accuracy, accuracy, readability, F1 score, confusion matrix, and the area under the rocking curve were obtained to evaluate the performance criteria.

The investigations indicated that logistic regression and random forest algorithms performed the best. Naive Bayes and decision tree algorithms need improvement.