A general methodology for data-based rule building and its application to natural disaster management
Risks derivedfromnaturaldisastershaveadeeperimpactthanthesoledamagesufferedbytheaffected zone anditspopulation.Becausedisasterscanaffectgeostrategicstabilityandinternationalsafety, developedcountriesinvestahugeamountoffundstomanagetheserisks.Alargeportionofthesefunds are channeledthroughUnitedNationsagenciesandinternationalnon-governmentalorganizations (NGOs),whichatthesametimearecarryingoutmoreandmorecomplexoperations.Forthesereasons, technologicalsupportfortheseactorsisrequired,allthemoresobecausetheglobaleconomiccrisisis placingemphasisontheneedforefficiencyandtransparencyinthemanagementof(relativelylimited) funds.Nevertheless,currentlyavailablesophisticatedtoolsfordisastermanagementdonotfitwellinto these contextsbecausetheirinfrastructurerequirementsusuallyexceedthecapabilitiesofsuch organizations.Inthispaper,ageneralmethodologyforinductiverulebuildingisdescribedandapplied to natural-disastermanagement.Theapplicationisadata-based,two-levelknowledgedecisionsupport system(DSS)prototypewhichprovidesdamageassessmentformultipledisasterscenariostosupport humanitarianNGOsinvolvedinresponsetonaturaldisasters.Avalidationprocessiscarriedoutto measuretheaccuracyofboththemethodologyandtheDSS.
Computational experiments with a validation process and a case study. For this, three inference methods with their own set of rules were created.
Build fuzzy rules from data with an inductive methodology. Rule aggregation and inference are then performed by means of a weighted averaging operators approach.
Predicted results about the number of casualties, number of injured, homeless, affected people and damage in US dollars by the SEDD.
Based on the discussion that current management DSS (DSS-DM) do not meet the needs of NGOs, the authors plan to tackle following problems:DSS-DM are not designed to address the specific problem of to any possible natural in any placeThe sophistication and infrastructure requirements of DSS-DM usually exceed those available in NGOs
No data was collected. Instead, the authors used data from the EM-DAT database ( Events Database), merged with UN data on the HDI (Human Development Index, used for estimating a country’s vulnerability)
Describe and apply a general methodology for inductive rule building for natural management. This application is a data-based, two-level knowledge Decision Support System (DSS) prototype which provides damage for multiple scenarios to support humanitarian NGOs involved in to natural disasters.
Although the system is a prototype, validation results suggest the suitability of the approach developed in this paperMore research with the SEDD project is neededThe authors emphasize the to develop decision support tools specifically to address the problem of unrealistic complexity in standard existing DSS. They show that it is possible to design such a practical decision support so that it can be implemented in context such as developing countries or NGOs.
During the validation process, one method classified more than two out of three instances of the validation set correctly. The other two methods didn´t perform that well, one of them performing rather poorly.The results in the case study were also quite poor. As in the validation process, the predicted values were very far away from the real ones.
Not applied as it is a computational
Presenting a new version of SEDD (Expert System for Diagnosis) which uses fuzzy logic. This SEDD is intended to be a web-available, low-cost, tailor-made that should fit specific NGO constraints such as ease of use, low computational and personnel requirements, and not relying on highly sophisticated and precise data.
DRIVER+ project has received funding from the European Union's 7th Framework Programme for Research,