GitHb repository to host python code (K_Means_Vectorized.py) for the K-means algorithms. Algorithms used can be found in "Information Theory, Inference & Learning Algorithms" by David J.C. Mackay

This repository also contains a Jupyter Notebook (Lab Session k-means Clustering Wine Chemical Composition Data.ipynb) which applies K-means algorithms using the Sci-Kit Learn Python Library. This notebook has been used for intereactive MSc Data Science undergraduate teaching.

A number of improvements could be made to this code and repository. Examples include:

13. Implement Updated Soft K-means Version 1 algorithm
14. Implement Soft K-means Version 2 algorithm
15. Implement Soft K-means Version 3 algorithm
16. Implement the silhouette method

Completed improvements include:

1. Add a feature to stop 'update loop' once mean values do not move
2. Reduce the amount of print statements by using loops and a python dictionary
3. Expand codes to work with more than 3 dimesnions
4. Add option for 2D functionality using an 'if' loop and additional input parameter
5. Split code into functions which may also easy implementation of elbow method
6. Improve the test data so that it contains 5 dimensions (Moved to OLD_CODE)
7. Implement elbow method
8. Improve the test data so that it contains 10 dimensions (Moved to OLD_CODE)
9. Change code to allow automated elbow method
10. Vectorize the distance and update calculation
11. Add in statistical running capability testing different random number seeds
12. Adhere to PEP8 Python standard

The 'wine' dataset is available at: https://archive.ics.uci.edu/ml/datasets.php