Skip to Main content Skip to Navigation

Informational Measures of Aggregation for Complex Systems Analysis

Robin Lamarche-Perrin 1 Jean-Marc Vincent 2 Yves Demazeau 3
2 MESCAL - Middleware efficiently scalable
Inria Grenoble - Rhône-Alpes, LIG [2007-2015] - Laboratoire d'Informatique de Grenoble [2007-2015]
LIG [2007-2015] - Laboratoire d'Informatique de Grenoble [2007-2015]
Abstract : The analysis of systems' dynamics lies on the collection and the description of events. In order to scale-up classical analysis methods, this report is interested in the reduction of descriptional complexity by aggregating events' properties. Shannon entropy appears to be an adequate complexity measure regarding the aggregation process. Some other informational measures are proposed to evaluate the qualities of aggregations: entropy gain, information loss, divergence, etc. These measures are applied to the evaluation of geographic aggregations in the context of news analysis. They allow determining which abstractions one should prefer depending on the task to perform.
Complete list of metadatas
Contributor : Arnaud Legrand <>
Submitted on : Wednesday, February 13, 2013 - 3:03:26 PM
Last modification on : Friday, July 17, 2020 - 11:10:29 AM


  • HAL Id : hal-00788019, version 1



Robin Lamarche-Perrin, Jean-Marc Vincent, Yves Demazeau. Informational Measures of Aggregation for Complex Systems Analysis. [Research Report] RR-LIG-026, 2012, pp.21. ⟨hal-00788019⟩



Record views