A topology-preserving system for environmental models forecasting

Abstract

This inter-disciplinary study presents a novel mathematical simulation model based on an algorithm for the summarization of self-organizing maps ensembles applied under the case-based reasoning (CBR) methodology to perform forecasting tasks. This methodology represents a knowledge-extraction frame, where past information is used to generate new solutions to new problems. The novel summarization algorithm based on topology-preserving models organizes the stored information simplifying the retrieval of the most useful information from the case base. This algorithm is used to organize the case base and to improve the speed and efficiency of the retrieval phase of the CBR cycle within the explained predicting system. The developed mathematical system was applied to a real case of study: a forest fire forecasting data set. Forest fires represent an environmental risk that should be predicted in order to avoid further damages. This novel system was able to predict the future situation of geographic areas after a forest fire had been originated.

Keywords:

• case-based reasoning
• forest fires
• topology-preserving models
• self-organizing maps
• ensemble fusion
• mathematical simulation

2000 AMS Subject Classifications :

• 68U35
• 68T05
• 68T10
• 68T30
• 92B20

Acknowledgements

This research has been partially supported through projects BU006A08 and SA071A08 of the JCyL and project CIT-020000-2008-2 of the Spanish Ministry of Education and Innovation. The authors would also like to thank the vehicle interior manufacturer, Grupo Antolin Ingenieria, S.A., within the framework of the project MAGNO2008-1028-CENIT Project funded by the Spanish Ministry.