The success of brain tumor treatment depends on a prompt and precise detection of a tumor. The rate of success in neurosurgery is continually rising, thanks to the ongoing advancement of technology that reduce the risk of problems. Brain tumor are extremely difficult to identify because of numerous irregularities in tumor size and location. Magnetic resonance imaging (MRI) analysis also necessitates the expertise of a neurosurgeon. It might be difficult and time-consuming to identify tumor from MRI in developing countries due to a lack of skilled clinicians and understanding regarding tumor. The issue can be addressed by an automated approach. Brain tumor classification can be achieved by a variety of methods and algorithms. The primary goal of this research was to find the most effective feature extraction and Machine Learning (ML) model for classification, suggesting one that can learn automatically through training and then make a sound conclusion. To do this, we use a Grey Level CoOccurrence Matrix (GLCM) and a Local Binary Pattern (LBP) as feature extraction methods. Next, an ML model, such as KNN (K Nearest Neighbor), SVM (Support Vector Machine), or NB (Naive Bayes), is selected. Several metrics were used to compare the effectiveness of different combination of two feature extraction methods and three ML models for tumor classification. The raw MRI data is used by the model during the classification procedure to determine the various tumor types. We examined the accuracy of 6 different combinations and discovered that GLCM plus NB algorithm yielded the highest accuracy (94.96%)

Brain Tumor, Resize, Image Enhancement, Feature Extraction, Evaluation Metrics