I'm honored to report that our team won the Euroclear Text Mining Hackathon that took place between 17-18 Jun in Brussels. This hackathon focused on extracting key financial information buried in thousands of legal documents. The hackathon was organized by the European Data Innovation Hub who announced Teradata as a winner on their Tweeter feed.
Our team was thrilled with the win, especially since competition was fierce. Other teams were composed of IBM Watson, Anatella/Timi, Hadoop Scala and several teams with “cool kids who code in Python”. We were thankful for the opportunity to compete with such talented experts and distinguished technologies.
To secure the win, our team used an elegant blend of machine learning algorithms together with in-depth text analysis and feature extraction. On the technology side, the team used the Aster data science platform to deliver superior results in record time. Aster’s built-in performance enabled Teradata to submit high quality results way ahead of all the other teams in both the training and final test periods.
Our team also used Aster to stay in first position (place) throughout the hackathon, except for a couple of hours at dawn when the “mighty Python” team delivered a burst of creativity. After regaining the lead, we blazed to the final and won.
We are thankful to Euroclear and the European Data Innovation Hub for organizing this hackathon.