less than 1 minute read

Hierarchical clustering, K-Means Clustering, T-Confidence Intervals.

Hierarchical clustering is just another way to get an overview of our data during the exploratory phase, usually done by constructing a dendogram. We can couple these with heatmaps to get a matrix like visualization.

K-Means covers the same purpose of clustering our data and providing a better idea of our higher dimensional data. This works by simply guessing centroids of clusters and then assigning points to centroids and then updating the centroids as average of the assigned points and then iterating.

T-tests:

  • paired
  • independent