Unsupervised Learning

Unsupervised Learning!

Principle components analysis

Clustering

K-means Clustering

Algorithm:

  1. Assign data points to a group
  2. Iterate through until no change:
    1. Find centroid
    2. Find the point that is closest to the centroids. Assign that data point to the corresponding group of the centroids.

How Many Groups

The art of chosing K.

Hierarchical Clustering

Bottom-up hierarchical groups can be read out from the dendrogram.

Published: by ;

Current Ref:

  • wiki/machine-learning/unsupervised/overview.md