The Nigeria Customs Service (NCS) has adopted new data analysis techniques to fight undervaluation as it has commenced analysing real-time data on imports coming through Tin Can Island Port, Lagos, using the Machine Learning (ML) algorithms to detect fraud.
According to information, the new data analysis techniques and ML were developed by the WCO BACUDA Team. BACUDA is a collaborative research project between Customs and data scientists. Its objective is to develop data analytics methodologies as a response to World Customs Organisation members’ requests for assistance in this domain, which is also one of the priorities defined in the Organization’s current Strategic Plan.
To develop the algorithms, BACUDA analysts used Customs data at the most disaggregated level, i.e. the transaction level. Such data was collected from customs administrations who wished to support the project including Nigeria Customs which provided a subset of five-year import data to enable the BACUDA Team experts to develop and test the algorithms for detecting fraud.
The information noted that the results obtained from analysing the data on imports coming through Tin Can Island Port using the BACUDA algorithms will be compared with those obtained from analysing the data on imports coming through Onne Port using traditional analytical methods.
The WCO Deputy Secretary General, Ricardo Treviño Chapa, who attended the pilot project on 2 March 2020, expressed confidence that the methods developed by the WCO would enable the Nigeria Customs Service to detect under-valuation with greater accuracy and, by so doing, to facilitate operations by compliant traders, create a fair business environment and increase revenue collection.
Mr. Treviño Chapa also appreciated the Nigeria Customs Service for the support provided to the BACUDA Project and for showing its strong commitment towards the adoption of modern targeting techniques.