Extensión del software para la visualización de grafos Gephi con algoritmos de agrupamiento asistidos por el usuario
Archivos
Fecha
2012-07-03
Autores
González Moreno, Diana Rosa
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Central “Marta Abreu” de Las Villas
Resumen
El análisis de grupos es útil en varios tipos de análisis exploratorios de patrones, en la toma
de decisiones, en situaciones de aprendizaje automático, incluyendo la minería de datos.
Muchos métodos de agrupamiento basados en las relaciones de los objetos representados en
un grafo tienen un alto costo computacional. El algoritmo de agrupamiento basado en la
intermediación diferencial fue creado para manejar grafos ponderados y no ponderados;
posee una alta complejidad temporal y necesita de parámetros tales como el tamaño de la
vecindad, que influencia esta complejidad temporal. Existen numerosas herramientas para
la visualización y manipulación de grafos, entre ellas Gephi. El objetivo de esta
investigación consiste en integrar el algoritmo de intermediación diferencial asistido por el
usuario en el software para la visualización de grafos Gephi. Los resultados obtenidos
muestran que integrar el algoritmo de agrupamiento basado en la intermediación diferencial
en Gephi puede resultar ventajoso, pues le permite al usuario predecir mejor los valores
óptimos de los parámetros y se pueden observar gráficamente los resultados del algoritmo.
The analysis of groups is useful in several types of exploratory analysis of patterns, e.g. in decisions making, in situations of automatic learning, including data mining. Several clustering methods based on the relations of the objects represented in a graph have high computational cost. The clustering algorithm based on the differential betweenes was created to manage pondered and not pondered graphs; it has a high temporary complexity and needs some parameters such as the size of the vicinity, which influences its temporary complexity. Numerous tools exist for the visualization and manipulation of graphs, among them Gephi. The general objective of this investigation is the integration of the differential betwennes algorithm with user interaction in the graph visualization software Gephi. The obtained results show that the integration of the clustering algorithm based on the differential betwennes into Gephi can be advantageous, because it allows the user to better predict good values of the parameters and the results of the algorithm can be graphically observed.
The analysis of groups is useful in several types of exploratory analysis of patterns, e.g. in decisions making, in situations of automatic learning, including data mining. Several clustering methods based on the relations of the objects represented in a graph have high computational cost. The clustering algorithm based on the differential betweenes was created to manage pondered and not pondered graphs; it has a high temporary complexity and needs some parameters such as the size of the vicinity, which influences its temporary complexity. Numerous tools exist for the visualization and manipulation of graphs, among them Gephi. The general objective of this investigation is the integration of the differential betwennes algorithm with user interaction in the graph visualization software Gephi. The obtained results show that the integration of the clustering algorithm based on the differential betwennes into Gephi can be advantageous, because it allows the user to better predict good values of the parameters and the results of the algorithm can be graphically observed.
Descripción
Palabras clave
Software de Visualización, Metodología, Grafos Gephi, Algoritmos de Agrupamiento, Usuarios