Compared with the standard stationary trade the online trade is an eldorado for small and big “swindlers” who may even operate as well organized criminal gangs. The rule “goods against money” cannot be directly one to one realized in the online trade, like in other branches of mail order business, except for cash on delivery. Clearly, to offer cash on delivery as single method of payment for a fraud secure online trade is an obstacle for this flexible type of trade. For this reason, in online shops a variety of methods of payment can be found, from tendering of account (usually restricted to a certain amount) over debit to customer or credit card.


However the problem is: How can the trader recognize if a person who makes an order is a customer who will eventually pay the goods?

In this year the DMC Contest dealed with the thematic invention, how Data Mining can ascertain the risk of loss of payments and thus, how Data Mining can reduce this risk.

For the participants it was a matter of generate a Data Mining model to classify orders in the classes “High Risk” or “Low Risk”.




1st Place:

Ilja Bezrukov, RWTH Aachen

Winner of DATA MINING CUP 2005

2nd Place:

Claudine Groß, Universität Karlsruhe

3rd Place:

Markus Weber, Universität Karlsruhe

4th Place:

Claudia Marino, Universität Karlsruhe

5th Place:

Volker Buege, Universität Karlsruhe

© 2022 GK Artificial Intelligence for Retail AG