U-control chart-based differential evolution clustering for determining the number of clusters in k-Means

Data are critical sources in several organizations and therefore the efficiency of access to it, sharing, extracting information and making use of that information has become an urgent necessity. Currently, data mining is one of the most recognized fields of research and application for carrying out such tasks. In general terms, data mining is the process of extracting useful information, previously unknown patterns and trends, from large databases. The data mining has received a great impulse in the last times motivated by different causes: (a) the development of efficient and robust algorithms for the processing of large volumes of data, (b) a cheaper computational power that allows the use of computationally intensive methods, and (c) the commercial and scientific advantages that have offered this type of techniques in the most diverse areas (Silva et al in Procedia Comput Sci 151:1219–1224 [1], Dianne and Deborah in Interactive and dynamic graphics for data analysis: with R and Gobi [2]). This paper presents an evaluation from different perspectives of a number of relevant nonsupervised quality measures.

• Autores: Rondón, Carlos; Romero-Pérez, Ivon; García Guliany, Jesús; Steffens Sanabria, Ernesto
• Tipo de recurso: Article of investigation
• Fecha de publicación: 2019
• Institución: Corporación Universidad de la Costa
• Repositorio: REDICUC - Repositorio CUC
• Idioma: eng
• OAI Identifier: oai:repositorio.cuc.edu.co:11323/10884
• Acceso en línea: https://hdl.handle.net/11323/10884; https://repositorio.cuc.edu.co
• Palabra clave: Clustering; K-means; Nonsupervised measures; Particle swarm optimization
• Rights: closedAccess
• License: Atribución-NoComercial-SinDerivadas 4.0 Internacional (CC BY-NC-ND 4.0)