Abstract: Disclosed are various embodiments for the identification and assignment of classifications for items. One of a plurality of classifications are identified in at least one computing device based at least in part on a plurality of tokens associated with the item and a data dictionary. An identified classification is thereby produced, and the identified classification is associated with a confidence score. The identified classification is automatically assigned to the item when the confidence score meets a threshold. A manual confirmation of the identified classification is obtained when the confidence score is below the threshold. The manual confirmation indicates whether the identified classification is erroneous.

Abstract: Systems and methods are provided for classifying a new item for purposes of estimating import fees associated with shipping the item internationally. In some embodiments, the classification is determined by calculating a similarity score for each of a plurality of previously classified items, each similarity score representing the similarity between a given previously classified item and the new item, identifying a similar item match for the new item, the similar item match corresponding to the previously classified item having the highest similarity score among the plurality of previously classified items, and assigning an item classification code associated with the similar item match to the new item.