Caracterización de flujos de datos usando algoritmos de agrupamiento

• Main Authors: Fabián Andrés Giraldo, Elizabeth León, Jonatan Gómez
• Format: Article
• Language: Spanish
• Published: Universidad Distrital Francisco Jose de Caldas 2013-09-01
• Series: Tecnura
• Subjects: clustering methods; data stream; data mining
• Online Access: http://tecnura.udistrital.edu.co/ojs/index.php/revista/article/view/635/569

This paper presents introductory materials to data-stream mining processes using clustering techniques. The limitations of traditional techniques are observed and the various approaches found in the literature are explained. The major trends in the different algorithms indicate that most applications separate the process into two phases; namely an online phase, which makes a data stream summarization in addition to the application of decay functions regardless of the data, and an offline phase, which is the application of traditional clustering techniques in order to obtain the cluster requested by users.The net result of this paper is a selection of desirable characteristics of an algorithm, based on the theoretical underpinnings of each of the works analyzed.