An Alzheimer’s Disease Category Progression Sub-grouping Analysis Using Manifold Learning On Adni

This study explores the use of autoencoders as an alternate feature embedding to PCA for unsupervised AD stage segmentation. Different manifolds are analysed to assess their value in analysing AD progression categories using the ADNI dataset. The learnt embeddings are visualised and a clustering assessment using DBSCAN is conducted. Classification accuracy is then assessed using the learnt embeddings and a Support Vector Machine (SVM).