EVALUATION OF INITIALIZATION METHODS FOR THE PERFORMANCE OF THE K-MEANS ALGORITHM

The K-means algorithm is one of the most widely used unsupervised machine learning methods; it helps to sort data clusters into a given number of groups with a pattern association that identifies relevant information in research domains. The heuristics of the algorithm are adaptable and easy to implement; however, one of its most notorious weaknesses is the poor assignment of K groups. This paper aims to analyze the different means of initialization and performance of the algorithm, as well as some applications of K-means in different industry sectors through a literature review, addressing relevant aspects to conclude in which cases transcendent results are obtained. Keywords: K-means, heuristic, clustering, performance, initialization, algorithm, centroids, metaheuristic, accuracy, evaluation, methods, applications.