An Algebraic Approach to Constrained K-means Clustering
An algebraic approach to k-means clustering with pairwise constraints is proposed, which simplifies the constraints using Gr\"{o}bner bases before optimization, improving the efficiency of cluster assignment and detecting whether the constraint set has a feasible assignment before optimization.
Hiroki Oohori, Kentaro Fukushima, Michio Yamamoto
· 0 citations