HARMONIZED SYSTEM CODE RECOMMENDATION: A MULTI-CLASS CLASSIFICATION MODEL

Evaluating the accuracy of Harmonized System (HS) code for goods manually is both error-prone and time-consuming. The limitation of capable customs officers with adequate knowledge in classifying goods, also the increase of cross-border transactions due to the new concept of e-commerce make customs officials face great challenges to meet its duty. This paper explored an automated approach employing machine learning algorithms to build the most suitable model to verify the correctness of description according to its HS Code. This project used historical data from declared import documents and official assessments from customs officers. Amongst the explored model, it is proved that Linear Support Vector Classification held the highest accuracy to classify HS Code based on its description. This findings contribute to the body of knowledge by providing a practical solution that leverages machine learning for HS code classification. The study's implications extend to enhancing customs operations and automation.