CDQ Data Quality as Service

Many users only search with name and country as search criteria. A suggested Golden Record based on such vague input has a high risk of being wrong. Therefore a Golden Record should only be generated if the results can be clearly related to the input, e.g.

• there was only one result (e.g. input: "CDQ GmbH, Germany"; result: only one record with full match for given input criteria)

• the given locality information clearly identifies a record from the results as base record. This could be city, postal code, street, house number dependent on the results (e.g. input: "BASF, Ludwigshafen"; result: only one record with name BASF in Ludwigshafen while "BASF, Germany" would lead to many possible full matches)

• an identifier was given

Legal forms should be extracted from input and results before comparison. Given legal forms should have an impact on matching score (only compare technical keys because of different abbreviations etc.) but if no legal form was given only names should be compared.