2015-07, Josh Montague
An overview of some approaches to selecting the 'k' (ie the number of clusters) when implementing k-means, the unsupervised learning algorithm.
This session was built using:
- Python 2.7.9
- IPython 3.0
- numpy 1.9
- scipy 0.15
- matplotlib 1.4
- seaborn 0.5
2017-05, update:
This notebook should run with versions:
- Python 3.6
- Juptyer 1.0
- scikit-learn 0.18
- numpy 1.12
- scipy 0.19
- matplotlib 2.0
- seaborn 0.7
Enjoy!