Classifying cancer subtypes from gene expression data. The work combines unsupervised techniques — clustering and dimensionality reduction — with supervised models including decision trees and random forests, applied to the TCGA Pan-Cancer dataset.
A fuller write-up is still to come.