Alzheimer’s disease is a neurodegenerative condition where neurons within the brain stop functioning, lose connection with other neurons and die. It is the most common cause of dementia, a loss of brain function that can adversely impact memory, thinking, language, judgment and behavior. There is no cure for this disorder but physicians can diagnose it at a very early stage, slow down the development, and improve patient's overall life quality. According to the WHO, Alzheimer's is a disease of affluence, which means their will be a greater demand for neuro specialists in the future.
This notebook aims to evaluate the performance of machine learning models in detecting and classifying Alzheimer's disease using MRI scan data of patient's brains. To complete this task, six diverse predictive models were built in scikit-learn and TensorFlow libraries:
- support vector machines,
- random forest,
- k-nearest neighbors,
- EfficientNetB0,
- InceptionV3,
- customized CNN (architecture from scratch).
We achieved such accuracies in model evaluation. The results of this project prove that the use of MRI scanning is very effective in identification of Alzheimer's disease. Applying machine learning in MRI image recognition has very high potential and looks promising for supporting physicians diagnoses.