The Border K-Means Clustering Algorithm for One Dimensional Data

Abstract

Clustering has been widely used for data pre-processing, mining, and analysis. The k-means clustering algorithm is commonly used because of its simplicity and flexibility to work in many real-life applications and services. Despite being commonly used, the k-means algorithm suffers from non-deterministic results and run times that greatly vary depending on the initial selection of cluster centroids. To improve the k-means algorithm, we present in this paper a border k-means clustering algorithm. It combines concepts from the k-means algorithm with an additional focus on the concepts of the borders dividing clusters. Consequently, the resulting border k-means algorithm leads to deterministic results and a great reduction in run time when compared with the traditional k-means algorithm.