Model Training
Train New Model

Evaluation uses a temporal holdout: models train on older events and are scored on the most recent ones — the same direction they're used in practice.

Model Library (13)
ModelTypeSamplesR2Trained
model_20260710_045316 lightgbm 20000 0.801 2026-07-10 04:53
model_20260710_030348 lightgbm 5000 0.822 2026-07-10 03:03
model_20260709_055343 lightgbm 5000 0.838 2026-07-09 05:53
model_20260709_034954 lightgbm 5000 0.852 2026-07-09 03:49
model_20260709_033235 lightgbm 5000 0.870 2026-07-09 03:32
model_20260708_170926 lightgbm 5000 0.849 2026-07-08 17:09
model_20260708_160629 lightgbm 5000 0.862 2026-07-08 16:06
model_20260708_152321 lightgbm 5000 0.856 2026-07-08 15:23
model_20260708_070531 lightgbm 5000 0.832 2026-07-08 07:05
model_20260708_023940 lightgbm 5000 0.863 2026-07-08 02:39
model_20260707_161728 lightgbm 1000 0.771 2026-07-07 16:17
model_20260706_021028 lightgbm 500 0.688 2026-07-06 02:10
model_20260704_123320 lightgbm 500 0.709 2026-07-04 12:33
Latest Training Results — 2026-07-10 04:53
20000
Samples
lightgbm
Model
0.78s
Runtime
Fill Rate Model (own-event demand + venue history) 57% better than baseline
0.801
R-squared
0.040
MAE
0.051
RMSE
0.095
Baseline MAE
marketing_cost
1071
ticket_price
966
dayofyear
769
venue_prior_event_count
632
duration_hours
547
entertainment_cost
466
venue_prior_mean_fill
343
max_capacity
269
Good model performance (R-squared > 0.7).
Training History (13)
DateModelSamplesR2StatusRuntime
2026-07-10 04:53 lightgbm 20000 0.801 OK 0.78s
2026-07-10 03:03 lightgbm 5000 0.822 OK 0.39s
2026-07-09 05:53 lightgbm 5000 0.838 OK 0.41s
2026-07-09 03:49 lightgbm 5000 0.852 OK 0.4s
2026-07-09 03:32 lightgbm 5000 0.870 OK 0.36s
2026-07-08 17:09 lightgbm 5000 0.849 OK 0.39s
2026-07-08 16:06 lightgbm 5000 0.862 OK 0.36s
2026-07-08 15:23 lightgbm 5000 0.856 OK 0.33s
2026-07-08 07:05 lightgbm 5000 0.832 OK 0.31s
2026-07-08 02:39 lightgbm 5000 0.863 OK 0.33s
2026-07-07 16:17 lightgbm 1000 0.771 OK 0.19s
2026-07-06 02:10 lightgbm 500 0.688 OK 0.15s
2026-07-04 12:33 lightgbm 500 0.709 OK 0.17s