Checklist NEOAI 2026 Day 1 · Task 3
Kaggleforces
Predict, from historical Kaggle results, which participants will finish in the top 3% of an upcoming Featured competition.
The task
The data are historical results of Kaggle Featured competitions: public and private leaderboard ranks and scores, numbers of submissions, medals, reward types and amounts, and the optimisation direction of each competition's metric. For every (competition, user) pair in the test competitions the contestant must output a continuous score that is higher for users more likely to finish in the top 3%.
The target is PrivateLeaderboardRank / TotalTeams ≤ 0.03, always computed relative to the full original leaderboard. Test users were filtered (at least 4 competitions and at least 10 total submissions), and the history pool contains only competitions that started before the earliest test competition.
Abridged by SOTA from the official materials. The official statement has the exact rules, and it wins wherever this summary differs.
At a glance
- You get
train.parquet(UserId, CompetitionId, HostSegmentTitle, EnabledDate, DeadlineDate, TotalTeams, PublicLeaderboardRank, PrivateLeaderboardRank,final_public_score,final_private_score,n_submissions, Medal, RewardType, RewardQuantity, EvaluationAlgorithmIsMax);test.csv(CompetitionId, UserId, TeamId);sample_submission.csv.- You submit
- CSV with columns Id (<CompetitionId>_<UserId>) and
pred_score(continuous). - Scoring
- ROC-AUC between the predicted score and the observed top-3% target.
- Rules
- General NEOAI 2026 rules apply.
- Format
- NEOAI 2026 main round, Day 1 (3 May 2026, 07:00–13:00 UTC). Individual.