Application of Business Intelligence for Customs Declaration: A Case Study in Indonesia

Abstract

The customs declaration made by the self-assessment needs to be re-examined by the document inspector. Customs declaration may have potential errors to determine even intentionally or not. However, the re-examination by the document inspector has not been optimal. This study uses business intelligence to address this problem by providing analysis capabilities for document inspectors. This research focuses on developing a data warehouse with the Kimball methodology. The results of research in the form of data warehouse design. The data warehouse will be used later for dashboards, OLAP, and data mining in order to detect fraud. The data mining algorithms used are firmness tree, support vector machine, neural network, and several ensemble methods. The results of the data mining process show that SMOTE has a significantly increased sensitivity score than non-SMOTE technique.