My randomised forests and I won both parts (easy and hard) of the undergraduate track of the UCSD Data Mining Contest 2009!
The contest is known for using real world data, provided by FICO. The contest is organised annually by the UC San Diego. 301 Teams from all over the world tried their best in detecting fraudulent anomalies in e-commerce data.
Six weeks of hard work paid off: nine students of our lab-team took places 1 to 8 in the top 10 of this year's Data-Mining-Cup. 688 participants from 159 universities from all over the world produced 248 solutions.
My submission — the third place — got lost somewhere during evaluation, so the third place was subsequently awarded twice.
My team (Pavlo Golik, Jan Henrik Ziegeldorf and me) placed second in a competition for the development of a computer player that was able to play a modified version of the board game Ingenious.
The competition consisted of ten teams and was funded by Sun Microsystems and organised by the Chair for Algorithms and Complexity.
I finished fifth in a competition dealing with predicting consumer behaviour, supported by eight other participants organised in a lab course of the Chair for Human Language Technology and Pattern Recognition.
688 students from 159 different countries participated in this award, founded by Prudsys AG.
My team (Jan-Thorsten Peter, Jan Henrik Ziegeldorf and I) won the competition, implementing and extending a snapshot algorithm.
The award was funded by Sun Microsystems and Deutsche Bank and organized by the German Gesellschaft für Informatik (Association for Computer Science).