Jun 13, 2013 — K-Nearest Neighbor Clustering (KNN) — K nearest neighbor (KNN) clustering is a supervised machine learning method that predicts a class label based on looking at other labels from the dataset that are most similar.  In a nutshell, the only things that you need for KNN... Read more

Jun 13, 2013 — Linear Regression — What is Linear Regression? What does a dataset look like? How do we fit a model? Batch Gradient Descent Stochastic Gradient Descent The Normal Equations Summary What is Linear Regression? Linear regression is modeling some linear relationship between a dependent... Read more

Jun 12, 2013 — FDR Correction — The False Discovery Rate (FDR) is a method used to correct for multiple comparisons in a statistical test. It is abstractly a method that gives us the idea of the false positive rate not for one single hypothesis, but for... Read more

Mar 31, 2013 — Brain Message 1.0 to Write Secret Messages into fMRI — Have you ever seen interesting shapes emerge as components when running independent component analysis (ICA) on resting BOLD functional MRI (fMRI) data? What comes to mind is the one time I swear that I saw the Koolaid Pitcher mascot, and... Read more