The results of the second round of the Data Clustering Contest are in.

In our evaluation, the Stage 2 tasks – indexing and ranking – had the greatest influence. Clustering tasks similar to the previous round (sorting by language, news or not, categories and threads) carried a lower weight. Overall speed of the algorithms influenced the final score.

See our comments on contest.com for more information on each of the submissions.

You can also view the output for each of the submissions for three different days (evaluation data sets: 1, 2, 3) and see how they work in real time in the Today tab.

In addition to the main prizes, we are awarding bonus prizes to those solutions that delivered significantly better results than the rest of the submissions in each of the tasks. If a submission required adjustments or fixes to make it work, it was penalized accordingly.

1st PLACE

🥇 Mindful Squirrel – €17,000 + €4,800 bonus
Bonus prizes for:
- News, Russian
- Categories, English
- Categories, Russian
- Threads, English
- Threads, Russian
Penalties (-€200):
- Required rebuilding from source code

2nd PLACE

🥈 Daring Frog – €12,000 + €5,500 bonus
Bonus prizes for:
- Language detection
- News, English
- News, Russian
- Categories, Russian
- Threads, English
- Threads, Russian
Penalties (-€500):
- Algorithm can't process simultaneous requests which makes it impossible to run the submission on real-time data

🥈 Swift Skunk – €12,000 + €900 bonus
Bonus prize for:
- Top news, English
Penalties (-€100):
- Algorithm required all articles to be in one folder

3rd PLACE

🥉 Gifted Lemur – €7,000 + €1,000 bonus
Bonus prize for:
- Threads, Russian

🥉 Ace Cock – €7,000 + €1,000 bonus
Bonus prize for:
- Top news, Russian

🥉 Mindful Kitten – €7,000

🥉 Happy Ladybird – €7,000

4th place

🏅 Hip Hyena – €4,000 + €1,000 bonus
Bonus prize for:
- Threads, Russian

🏅 Bossy Gnu – €4,000

🏅 Hairy Snail – €4,000

Several participants achieved some of the best results in individual tasks, but didn't get a high enough overall score to compete for the main prizes. They are receiving bonus awards for those parts of their algorithms in which their solutions were significantly better than the rest.

Bonus Awards

🎖 Bright Deer – €1,000
- Threads, Russian

🎖 Fancy Beetle – €1,000
- Categories, English

🎖 Large Crab – €1,000
- Language detection

🎖 Suave Penguin – €800
- News, English (€1,000)
- Required rebuilding from source code (-€200)

🎖 Sweet Beaver – €800
- Categories, English (€1,000)
- Required rebuilding from source code (-€200)

Stay tuned for more news about our contests.