An efficient <i>k</i> ‐modes algorithm for clustering categorical datasets

Mining clusters from data is an important endeavor in many applications. The $k$-means method a popular, efficient, and distribution-free approach for clustering numerical-valued data, but does not apply categorical-valued observations. $k$-modes addresses this lacuna by replacing the Euclidean with Hamming distance means modes objective function. We provide novel, computationally efficient implementation of $k$-modes, called OTQT. prove that OTQT finds updates to improve function are undetectable existing algorithms. Although slightly slower per iteration due algorithmic complexity, always more accurate almost faster (and only barely on some datasets) final optimum. Thus, we recommend as preferred, default algorithm optimization.