Generalized and Multiple-Queries-Oriented Privacy Budget Strategies in Differential Privacy via Convergent Series

Algorithm 1

One iteration of differentially private k-means clustering.

(1)	Partition the points into clusters with a set of centroids
(2)	for each of the cluster , do
(3)	Count the noisy number of points in as
(4)	Compute the noisy sum of points in as
(5)	Update a new centroid .
(6)	end for