The role of artificial intelligence in managing customs risk for Algerian customs

Abstract

Nowadays, the trend toward adopting artificial intelligence (AI) in customs risk management has increased across the globe, yet it may not be efficiently perform the required if the design risk management process is not rich in many features and inputs that are accurate and mathematically applicable. In this paper, we have designed a recommendation system for customs risk management at the level of Algeria’s customs whose current risk management system is still based on the discretion of customs agents, supported by certain internal laws and regulations that guide the decisions of Algerian customs agents. In the design of this recommendation system, we relied on the supervised machine learning where we used five different algorithms. These algorithms have resulted in close accuracy ranging from 97% to 99%. This model reduces the time taken to process different shipments and supports the decision-making process for customs inspectors. However, the current approach of risk management at the Algerian customs level requires greater depth, quality and accuracy at the input level in order to build a highly efficient customs risk management model.