A data mining framework to detect tariff code circumvention in Turkish customs database

---

Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Enformatik Enstitüsü, Bilişim Sistemleri Anabilim Dalı, Türkiye

Tezin Onay Tarihi: 2012

Öğrenci: BURCU BAŞTABAK

Danışman: TUĞBA TAŞKAYA TEMİZEL

Özet:

Customs and foreign trade regulations are made to regulate import and export activities. The majority of these regulations are applied on import procedures. The country of origin and the tariff code become important when determining the tax amount of the merchandise in importation. Anti-dumping duty is defined as a financial penalty, published by the Ministry of Economy, enforced for suspiciously low priced imports in order to protect the local industry from unfair competition. It is accrued according to tariff code and the country of origin. To avoid such an obligation in order to not to pay tax, a tariff code that is different from the original tariff code may be declared on the customs declaration which is called as "Tariff Code Circumvention". To identify such misdeclarations, a physical examination of the merchandise is required. However, with limited personnel resources, the physical examination of all imported merchandise is not possible. In this study, a data mining framework is developed on Turkish customs database in order to detect “Tariff Code Circumvention”. For this purpose, four types of products, which are the most circumvented goods in the Turkish customs, have been chosen. First, with the help of Risk Analysis Office, the significant features are identified. Then, Infogain algorithm is used for ranking these features. Finally, KNN algorithm is applied on the Turkish customs database in order to identify the circumvented goods automatically. The results show that the framework is able to find such circumvented goods successfully.