Home/AI/ML Notes/Deep Learning and Machine Learning/Clustering
⌂ Main Page
Deep Learning and Machine Learning

Clustering

K-means and friends: algorithms, initialization, and evaluation.

Clustering

Kmeans cluster initialization: 1) choose random k points, 2) kmeans++ (first point is random, choose rest by probability proportional to the squared distance away from a given point's nearest existing centroid - attempt to push the centroids as far from one another as possible), 3) naive sharding (deterministic method, quicker than stochastic, approximates the spread of initial centroids across the data space via the composite summation value)