Clustering high dimensional Electronic Patient Records

We reduce patient records from very high dimensional down to 2D using marginalized denoising autoencodrs and t-SNE, followed by hierarchical clustering. Ther…