Scores computed by CI from submissions in results/. The instructor’s hidden test set is scored separately — a public/hidden gap is the overfitting signal.
Seismic source classification (Chapter 3.5) — macro-F1, higher is better¶
No submissions yet — the first scored run posts here.
CO2 forecasting diagnostic (Chapter 4.10) — MASE, lower is better¶
This track is a diagnostic, not a grade: the holdout months are public NOAA data. Grading weight sits on the hidden synthetic track below.
No submissions yet — the first scored run posts here.
Hidden synthetic forecast (Chapter 4.10, graded) — MASE, lower is better¶
Scored in the instructor’s run against the private truth series; results are announced in class.
No submissions yet — the first scored run posts here.