Datasets:

ClarusC64
/

clinical-quad-site-performance-data-integrity-protocol-drift-power-loss-v0.1

trial_id string	site_id string	week int64	enrollment_rate_per_week float64	data_query_rate float64	missing_data_fraction float64	protocol_deviation_rate float64	visit_window_violations float64	endpoint_variance float64	stat_power_estimate float64	power_loss_next_60d int64	label_power_loss_next_60d int64
TRIAL_SPI01	S01	4	4.2	0.12	0.05	0.04	0.06	1.1	0.88	0	0
TRIAL_SPI01	S01	6	4	0.15	0.06	0.05	0.07	1.2	0.86	0	0
TRIAL_SPI01	S02	8	3.5	0.22	0.09	0.07	0.09	1.4	0.81	0	0
TRIAL_SPI01	S02	10	3.1	0.28	0.12	0.09	0.12	1.7	0.75	1	1
TRIAL_SPI01	S03	12	2.8	0.34	0.15	0.12	0.15	2	0.69	1	1
TRIAL_SPI02	S01	5	4.5	0.1	0.04	0.03	0.05	1	0.9	0	0
TRIAL_SPI02	S02	7	4.1	0.14	0.06	0.05	0.07	1.2	0.87	0	0
TRIAL_SPI02	S03	9	3.4	0.25	0.11	0.08	0.1	1.6	0.78	1	1
TRIAL_SPI02	S03	11	3	0.31	0.14	0.11	0.13	1.9	0.72	1	1
TRIAL_SPI02	S04	6	3.8	0.18	0.08	0.06	0.08	1.3	0.83	0	0

Clinical Quad Site Data Protocol Power Loss v0.1

Each row is a site week snapshot.

Core quad

Site performance
Data integrity
Protocol drift
Statistical power

Target

label_power_loss_next_60d

Files

data/train.csv
data/tester.csv
scorer.py

Evaluation

Run model on data/tester.csv
Return predictions row aligned
Score with scorer.py

License

MIT