06_clustering.md

Section 06: Clustering

Clustering enables us to group the rows of a data set together based on their similarities without knowing anything about the true class labels of the rows in the data set. Clustering is useful in market segmentation, anomaly detection, and other unsupervised learning tasks. Clustering is also useful as a data preprocessing step for supervised learning tasks.

Class Notes

• Overview of clustering techniques - Blackboard electronic reserves

• Introduction to Data Mining clustering notes

• EM clustering example

• H2o clustering example

Sample Quiz

Quiz Key

Assignment

Supplementary References

• Introduction to Statistical Learning
  Section 10.3

• Introduction to Data Mining
  Chapter 8

• My Quora answers regarding clustering

• Estimating the Number of Clusters in a Data Det via the Gap Statistic
  by Robert Tibshirani, Guenther Walther, and Trevor Hastie

• A Tutorial on Spectral Clustering

• K-means tutorial with numpy