Imputing missing values in nuclear safeguards evaluation by a 2-tuple computational model

Rosa Rodriguez, Da Ruan, Jun Liu, Alberto Calzada, Luis Martinez, Jan Wagemans

Research outputpeer-review

Abstract

Nuclear safeguards evaluation aims to verify that countries are misusing nuclear programs for nuclear weapons purposes. Experts from International Atomic Energy Agency (IAEA) evaluate indicators using several sources such as State declarations, on-site inspections, non-safeguards IAEA databases and other open sources. These evaluations are aggregated to obtain a global assessment, which are not precise because the available information is vague or too difficult to be understood. Moreover, experts cannot often assess all evaluations because of too many indicators or because of the lack of sufficient expertise about some of indicators. This results in many missing values in the evaluations that can change final results. In this contribution, we impute the missing values by a 2-tuple computational model that calculates a trust measure to express the reliability of the imputations.
Original languageEnglish
Title of host publicationArtificial Intelligence and Soft Computing
Place of PublicationHeidelberg, Germany
Pages202-209
StatePublished - Jun 2010
EventICAISC 2010 - 10th International Conference on Artificial Intelligence and Soft Computing - ICAISC, Zakopane
Duration: 13 Jun 201017 Jun 2010

Publication series

NameLecture Notes in Artificial Intelligence (LNAI)
Number6113

Conference

ConferenceICAISC 2010 - 10th International Conference on Artificial Intelligence and Soft Computing
Country/TerritoryPoland
CityZakopane
Period2010-06-132010-06-17

Cite this