Frequently Asked Questions

The DATA MINING CUP is an international student competition. Since the year 2000 GK Artificial Intelligence for Retail AG (formerly prudsys AG) has organized the tournament for student teams. Every year we are confronted with questions around the DATA MINING CUP – most of them should be answered below.

If your question does not get answered, please don’t hesitate to contact us.

Please also refer to our conditions for participation. They contain answers to some more questions and clarify certain points of the competition.

No, only students with a valid enrollment certificate can take part.

In this case contact one of the two team leaders; you will find their contact data in the team summary. Each team leader is obliged to include new team members who meet the participation criteria into the team.

Yes, with certain restrictions: From 2019 onwards students from each educational institution may be part of one additional mixed team. Mixed teams are built by students from a maximum of two educational institutions. All institutions must be specified at registration. Registration is only possible via e-mail with the organizer.

No. A team leader can only register and lead one team.

Please check first whether the maximum number of permitted teams per institution has been reached in the team summary . If not, then register a new team online. Only when you receive the separate confirmation email giving you your team ID is your team authorised for participation.

There are no restrictions. There must be at least one person (team leader).

Yes. A team may consist of just one person (team leader). But each team leader is obliged to add potential students as members of the team. This applies in particular if the educational institution has already registered two teams.

No, this is not necessary. Only if the team comes among the teams who win the prize does the team leader have to list the team members before the prize is awarded and provide evidence of the student status for both the leader and the team members.

Yes, you are always allowed to upload a new solution, but only the last solution is accounted for.

No, all models are evaluated after the deadline on June 29, 2021. This is to prevent exploiting the competition by submitting exploratory solutions to improve the score.

Right now we are not disclosing which descriptive features are shown on the evaluation page, other than the information is part of the “items.csv” of the competition’s data set.

The evaluation page will go live on July 1. and will be open for votes for fourteen days until the 14th of July.

With every visit of the evaluation page a reference product from the evaluation set is drawn and together with three of the submitted solutions. The reference product as well as the solutions are displayed on the evaluation page with their book cover and descriptive information. The page visitor can select suitable recommendations with a click.

No, only one recommendation per solution is displayed. When the submitted solutions are drawn, a number k between 1 and 5 is drawn as well and only the k-th recommendation of every solution is displayed on the page. This has two reasons. First you don’t want to everwhelm the page visitor with too many choices. Secondly, displaying only one of the five recommendations makes it much harder for the participants to recognize their own solution.

Yes, a preview as well as an instruction for the voting mechanism will be available shortly before the voting period.

The number of appearances on the evaluation page as well as the number of votes are stored for every team. The team with the highest ratio of votes/appearances wins.

The teams are drawn evenly. The current ranking is not considered as it would either give an unfair advantage or disadvantage in either way.

Yes all 5 recommendations  have to be distinct. The submissions will be tested for that.

No, we do not provide access to the bookcovers. However using additional information for this contest is not prohibited. Therefore, if you have access to a data base to legally aquire the book covers you are free to use them.

© 2022 GK Artificial Intelligence for Retail AG