Classification of MRI Images for Brain Tumor Patient Screening

Abstract

This research aims to create a model for screening brain tumor patients using MRI imaging data from www.kaggle.com. The data was gathered by Chakrabarty (2021) with a total of 253 images. The data was analyzed using the cross-industry standard process for data mining, then the performances of the classification models were compared. The results showed that random forest technique gave the best result for predicting the likelihood of brain tumors, with an accuracy of 76.31%. The F-measure was 73.48% with a sensitivity of 70.14% and a specificity of 82.69%. The data analysis results could be utilized to develop an information system for future patient screenings for brain tumors.

Keywords

• performance comparison
• image classification
• brain tumor