Diagnosis-of-Alzheimers

This Repository contains code and analysis results for diagnosis of Alzheimer's disease on the Oasis-3 dataset. An input image is classified in four classes(Mild-Demented,Moderate-Demented,Non-demented,Very-Mild-Demented) with the help of the ML model.
The Repository contains the following contents:
Augmented.ipynb: This script can be used for reducing imbalances present between different classes in our training dataset.
Alzheimer's_Diagnosis.ipynb: Contains the tuned ML model that is used for classification
Classification_report.ipynb:Script for generating confusion matrix
Best_perfromance_with_Adam.pdf:Contains the Epoch information,training and testing accuracy for the best perfromance achieved using the tuned ML-network with Adam optimizer. The testing dataset accuracy was found to be above 70% for every run.
Diagnosis_of_Alzheimers.docx:Contains a detailed report illustrating different approaches and the process of fine tuning the parameters before selection of final model.