HS code is a 6-digits commodity code. A commodity code is a sequence of numbers that describes a good according to its nature, type, and purpose. There are different types of commodity codes depending on their number of digits: HS codes, HTS codes, or TARIC codes for example. These codes are harmonized, and they are an essential element of global trade because they are necessary to fulfill several compliance requirements. Check out below frequently asked questions about commodity codes.
Q: What is the Harmonized System (HS)?
HS codes, accepted by the majority of nations around the world, are identification codes given to goods for use in international trade. The HS codes are administered by the World Customs Organization (WCO) and are internationally accepted for use by customs authorities and companies to identify goods.
Q: Which countries` nomenclatures are supported?
Eurora AI-based HS Code Allocation service enables allocating of 6-digit HS codes that are similar for all the countries around the world.
We support county-specific nomenclatures:
6, 8 and 10-digit codes for the EU (Combined Nomenclature for 27 countries)
6, 8 and 10-digit codes for GB
8 and 10-digit codes for the US (Harmonized Tariff Schedule)
8 - digit codes for UAE
Q: What is the difference between HS and HTS codes?
HS codes are universal while HTS codes are country-specific. They are both used to classify imported goods on a global scale.
An HS Code is a standardized number associated with a given commodity that identifies that product worldwide. It is a 6-digit number universal to all countries under the International Harmonized System. An HTS code ranges from 7 to 10 digits and provides country-specific product and tariff classification.
In a nutshell, the HTS code is only used in the USA and is the same as 10-digit TARIC in the EU.
Q: How does the Eurora HS classification work?
Eurora uses customized ensembles of machine learning models. The algorithms are based on the following methods:
- FastText ensembles using customer-specific input representations.
- Transformer-based deep neural network models.
The choice of model depends primarily
- on the volume and the variety of customer data
- customer's field of activity.
Eurora constantly has about 10 models, approaches, and ways of combining models in the applied research phase. The Eurora platform allows a fuzzy search. "Inaccurate product description" can mean that due to the customer constraints, the description has extra noise in it, the description has typos, or the description is written using synonyms. Eurora solution performs preprocessing which removes any irrelevant information from the description so that the decision can be made only from the informative parts. In addition, instead of correcting the input, the underlying model making the decision is taught to be robust concerning the typos, word cases, and the use of language and words by using context and synonym approximation. Context allows to still derive the correct estimation if some parts of the description are inaccurate. However, if the product description is inaccurate to the level where the description no longer describes the product or service itself (or where it is no longer differentiable from some other product/service category), all bets are off, unless the customer has passed the information to Eurora, denoting that because of reasons, the customer wants some specific descriptions to be handled as something or that the customer is not dealing/is only dealing in a specific range of products. For existing customers, it's the domain experts and data scientists that work relentlessly to make the HS code allocation as universal as possible. At the same time, we put daily effort into monitoring the quality of the estimations we give to them and improving AI continuously.