Abstract
In brain cancer diagnosis, the interpretation of classification model results is crucial. In this study, we present an algorithm designed to graphically explain the performance of classification models, including the Support Vector Classifier (SVC) and Random Forest for processing medical images related to brain cancer. The aim is to evaluate the performance of machine learning in the classification of three types of brain tumours. The method allows us to visualise the pixels that these techniques consider most relevant in the decision-making process of the referred models. The results obtained show a promising performance in understanding the relationships between the input pixels of the medical images and the resulting classifications, facilitating the interpretation of the results and increasing their reliability, contributing significantly to more informed and accurate clinical decision-making.
Translated title of the contribution | Exploring Classifier Behaviour: Support Vector and Random Forest Machines in Brain Cancer Diagnosis through Medical Imaging |
---|---|
Original language | Spanish |
Pages (from-to) | 528-538 |
Number of pages | 11 |
Journal | RISTI - Revista Iberica de Sistemas e Tecnologias de Informacao |
Volume | 2024 |
Issue number | E66 |
State | Published - 2024 |
Bibliographical note
Publisher Copyright:© 2024, Associacao Iberica de Sistemas e Tecnologias de Informacao. All rights reserved.