Having all product data organized by classified tables of normalized attributes and values, enables us to group products by their similarities. In IT systems products are considered similar when they are described the same way. In realty users can fully interchange products that fall under the same definition of Form-Fit-Function (FFF). The DeDuper provides the user with a powerful tool to find products that share the same Form-Fit-Function.

The major challenge is to provide users with effective and intuitive ways to easily define their own Fit-Form-Function on demand, and to return all products exactly fall under those FFF definitions.

The roll of the DeDuper is to find products that share the Form-Fit-Function definition and to enable the consolidation of products with high level of similarity.

Usually we use the DeDuper after the extraction of the technical attribute (using the Extractor). The user can use predefined similarity definitions (FFF style) or define new, ad-hoc definitions. There are four levels of similarity: Exact, Same, Similar and Alternative. The DeDuper enables users to consolidate products by their similarity level.

The technology behind the DeDuper consists of stochastic and probability algorithms.

The deliverables of the DeDuper are four groups of products with different similarity levels and list of consolidated products (key-mapping).