Using Classification Proposals for Commodity Codes

Objective

After completing this lesson, you will be able to classify products under commodity codes using classification proposals.

Classification Proposals

SAP GTS can propose commodity codes for product classification based on an algorithm provided on the SAP HANA Platform ("SAP HANA fuzzy search").

Using the product description transferred from the feeder system, this algorithm calculates the similarity of products still to be classified to products already classified under a particular commodity code. If at least one sufficiently similar product can be determined, its commodity code is proposed in the Classify Products app in the Proposed Number column.

Note

If a commercial description is maintained in a product master record, the system can use this text as an alternative to the product description transferred from the feeder system.

SAP GTS generates classification proposals in two programs that are supposed to run in the background at regular intervals. They can be scheduled using the Schedule Creation of Trade Classification Attributes - Fuzzy Search (GTS) and Schedule Creation of Trade Classification Proposals – Fuzzy Search (GTS) apps.

Configuration of Fuzzy Search

In Customizing for SAP GTS, fuzzy search can be activated for numbering schemes representing nomenclatures for commodity codes (SAP Global Trade Services, edition for SAP HANAGeneral SettingsNumbering SchemesCommodity CodesDefine Numbering Scheme for Commodity Codes).

After activation of fuzzy search, the search parameters have to be added.

Fuzzy Search Parameters

The Reference Country/Region Key field only matters, if commercial descriptions are maintained in the product master and if the nomenclature represented by the numbering scheme is used in several countries. In the EU, for example, French is one of the official languages in Belgium, France and Luxembourg. Commercial descriptions might thus vary depending on the Member State in which SAP GTS is used.

The number entered in the Exactness Fuzzy Search field determines the fuzziness value in fuzzy search. The closer the similarity between a source term and the target term, the higher the score calculated in the search. If the search should produce precise results for individual terms, a high fuzziness threshold is therefore required.

A product or a commercial description however can consist of more than one term. After computing the individual scores of the terms forming a description, the overall score for the description is calculated from the similarity of all its individual terms. The number in the Minimum Text Score field is the threshold for this overall score.

In the comparison, the system divides the number of matching search terms of a product or commercial description by the total number of search terms that constitute the description. Therefore, a third threshold has to be specified in the Accuracy Threshold field. The number entered in this field determines the minimum percentage of the terms that have to match in a product or commercial description.