A Comparison between Linear SVM and Logistic Regression Classifiers towards the Early Diagnosis of Alzheimer's disease

Abigail Copiaco, Nidhal Abdulaziz
2017 International Journal of Latest Research in Engineering and Technology (IJLRET) ||   unpublished
Alzheimer's disease (AD) is a tenacious neurodegenerative brain disorder that adversely influences its victims' memory. Due to its progressive property, early detection of the disorder is crucial for regular observation of the patient's condition. This research concentrates towards determining an optimum classification algorithm for an early and rapid diagnosis of AD. The proposed method constitutes for high efficiency through a Computer visionimage enhanced MRI brain scan where brain volume
more » ... ere brain volume features are extracted via the regional MRI measurementtechnique, fed separately to a Linear Support Vector Machineand a Logistic Regression classifier for result verification and method comparison. Effectiveness of the suggested method is then computed through the accuracy, sensitivity, specificity, and precision. For 100 MRI brain scans, 30 being the test set, results indicate 73% accuracy using the Linear Support Vector Machines method, while 50% accuracy is achieved with Logistic Regression. However, improvement is expected upon the application of the same technique in a larger sample set of brain scans.