GoldenBullet is a semi-automatic product classification system.

Product classification for warehouse catalogues is a labour-intensive, complicated, time-consuming and error-prone process that requires tremendous efforts. In frequent repeated cycles, millions of millions of products have to be classified with regard to multiple existing standards (e.g. UNSPSC, GPC, e-Class) Furthermore, each of these standards may contain several thousands of target classes or customer specific taxonomies.

In the project GoldenBullet we aim to realize an innovative semi-automatic workflow to be delivered as service with a 95% classification accuracy via an integrated web application. This workflow seamlessly combines a fully automated product classifier with accurate manual classification and a background learning process. Further industrial cooperation will be fostered by applying industry scenarios to the workflow.