suva007 / clustering

CONTAINS CLUSTERING ALGORITHMS

CLUSTERING ALGORITHMS

kmeans.ipynb :

• best for datasets that has centroids at mean position of clusters.
• contains graphical representation of how centroid position changes with change in mean and for different iterations
• contains ARI AND silhouette SCORE metric as well.

pnng-2.ipynb

• minimum spanning tree based clustering algorithm.
• improves time complexity of this mst based approach from O(n^2) to O(n^3/2).
• for detail discription refer fast approximate mst.pdf

agglomerative_algorithm.ipynb

• hierarichal clustering everything you need.

mst_divisive.ipynn

• mst based clustering on not very desirable dataset.

DETERMINISTIC KMEANS.ipynb

• deterministic initialization algorithm for K-means (DK-means) by exploring a set of probable centers through a constrained bi-partitioning approach. The proposed algorithm is compared with classical K-means with random initialization and improved K-means variants such as K-means++ and MinMax algorithms.
• for detail description refer Jothi2019_Article_DK-meansADeterministicK-meansC (1).pdf

DC.ipynb

• categorical + numerical data.
• clustering algorithm = agglomerative(hierarichal).

dc_heart.ipynb

• purely numerical high dimensional data.
• clustering algorithm = agglomerative(hierarichal).

dc_last.ipynb

• categorical data = ordinal + nominal.
• clustering algorithm = kmode.

dc_nursery_kmod.ipynb

• categorical data = ordinal + nominal.
• clustering algorithm = kmode.