Smart electronic governance in international trade using blockchain and Naïve Bayes’ based text classification to prevent import duty frauds

Exports and imports are the major activity of any country. At the time of bills of exports/ imports document filing, the importers/ exporters declare the item description and associated tariff head code. Every tariff head has its duty rate associated. The importers/ exporters may declare the item in a different tariff head to take the duty benefits. The assessment officials must match the item description with the item tariff head. However, due to the ever-changing tariff head, it is impossible to manually check and verify whether the item descriptions lie in the declared tariff head. Thus, gathering a comprehensive dataset consisting of globally available customs tariff heads and product description combinations is a difficult task. This paper utilizes blockchain technology to securely store and manages the training datasets and uses the collected datasets to train a Naïve Bayes classifier probabilistic machine learning algorithm. By combining blockchain technology for data integrity and transparency with the predictive power of the Naïve Bayes classifier, the proposed novel methodology offers a robust solution for preventing duty frauds and ensuring compliance with customs regulations.