PENERAPAN METODE DECISION TREE PADA TRANSAKSI EKSPOR BARANG UNTUK IDENTIFIKASI KESALAHAN DATA EKSPOR DI DIREKTORAT JENDERAL BEA DAN CUKAI

Abstract

Data mining is an analysis process to find a design from a big and various data that is widely applied in many sector which is applied in this present era. Directorate General of Customs and Excise (DGCE) as government institution in traffic of import and export trading always faces the big and various data. DGCE has a problem about data accuracy level since the using of self assessment in document filling especially in export commodity document. To find the mistake of identification design in export data, we use data mining method with classification of decision tree using algorithm C4.5. The using data in experiment is export commodity data in Main Service Office (MSO) Tanjung Priok during 2017 with the population of 22.000 commodity data consist of 10 various export commodities. By using sampling random simple technique, 30 samples data are taken and be treated using algorithm C4.5 to count the entrophy and gain values. From the calculation is obtained, the higher gain and foreign exchange attribute= 0.681233 with extreme data and normal data. The further experiment between the foreign exchange attribute correlation and other data attribute is obtained the high correlation level in compatibility HS_code attribute and commodity description with gain value= 0.591673 as the final output from the branch result of decision tree which is shown, so the mistake of design in export data is found and affected by two rules of data attribute, they are foreign exchange value and compatibility HS_code with the commodity description. The design which is found from the decision tree proves that the using of decision tree classification method with algorithm C4.5 is effective to find the mistake of design in export data at DGCE.