A basic customs mission is the fight against fraud and trafficking. The conditions for carrying out this task depend on the creation of economic problems as well as on the actions of the actors in charge of carrying it out. As part of the customs clearance process, in connexion with the growth of international trade, customs are now faced with a growing volume of goods. To restrict intrusive control, automated risk management is therefore necessary. In this paper , in order to recognise suspicious behaviour arising from customs regulation, we suggest an unsupervised classification method to derive information rules from a database of customs offences ... The intention is to apply the Apriori concept in customs procedures on the basis of repeated data-based reasons relating to customs offences in order to discover possible rules relating to the relationship between a customs process and an offence for the purpose of collecting information relating to the incidence of fraud. This mass of sometimes heterogeneous and complex data thus creates new needs that must be able to be fulfilled by methods of information extraction. Inevitably, the determination of infringements includes the accurate recognition of threats. It is an original approach to constructing association rules in two steps focused on data mining or data mining. First, look for frequent patterns (support > = minimum support) and then create association rules from the frequent patterns (Trust > = Minimum Trust). Three key association rules were illustrated in the simulations carried out: the forecasting rules, the targeting rules and the neutral rules, with the inclusion of a third predictor of the importance of the law, the lift test.
Author(s) Details
Bi Bolou Zehero
Institut National Polytechnique -
Houphouët Boigny, Yamoussoukro, Côte d’Ivoire and Direction Générale des
Douanes, République Côte d’Ivoire.
Pacôme Brou
Ecole Supérieure Africaine des
TIC- ESATIC, Abidjan-Treichville, Côte d’Ivoire.
Olivier Asseu
View Book :- https://bp.bookpi.org/index.php/bpi/catalog/book/299
No comments:
Post a Comment