Dataset Viewer
Auto-converted to Parquet Duplicate
trial_id
string
site_id
string
patient_id
string
day
int64
biomarker_level
float64
biomarker_pos
int64
subpop_flag
int64
dose_mg
int64
exposure_auc
int64
endpoint_change
float64
endpoint_drift_z
float64
signal_fail_next_30d
int64
label_signal_fail_next_30d
int64
TRIAL_BSE01
S01
P0001
14
1.2
0
0
50
3,800
-0.1
0.2
0
0
TRIAL_BSE01
S01
P0002
21
1.4
0
0
50
3,950
-0.08
0.3
0
0
TRIAL_BSE01
S02
P0003
28
2.8
1
1
50
4,100
-0.22
0.5
0
0
TRIAL_BSE01
S02
P0004
35
3.1
1
1
75
5,600
-0.34
1.1
1
1
TRIAL_BSE01
S03
P0005
42
3.4
1
1
75
5,450
-0.3
1
1
1
TRIAL_BSE02
S01
P0006
14
2.6
1
0
50
4,050
-0.12
0.4
0
0
TRIAL_BSE02
S01
P0007
21
2.9
1
0
75
5,750
-0.18
0.6
0
0
TRIAL_BSE02
S02
P0008
28
1.3
0
0
75
5,200
-0.05
0.2
0
0
TRIAL_BSE02
S02
P0009
35
3.2
1
1
100
7,200
-0.38
1.3
1
1
TRIAL_BSE02
S03
P0010
42
3.5
1
1
100
7,050
-0.36
1.2
1
1

Clinical Quad Biomarker Subpopulation Dose Endpoint Drift v0.1

Each row is a patient state snapshot.

Core quad

Biomarker status
Subpopulation flag
Dosing strategy
Endpoint drift

Target

label_signal_fail_next_30d

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

Downloads last month
19