Datasets:

ClarusC64
/

ABX-CS-002_pneumonia_resolution_lag-v0.1

row_id string	series_id string	day int64	host_model string	abx_drug string	abx_exposure_index float64	severity_index float64	temp_c float64	expected_temp_c float64	temp_nonresponse_vs_expected float64	spo2_percent int64	expected_spo2_percent int64	spo2_nonresponse_vs_expected int64	resp_rate_per_min int64	expected_resp_rate_per_min int64	rr_nonresponse_vs_expected int64	pneumonia_coherence_index float64	later_icu_escalation_or_treatment_failure_flag int64	assay_method string	source_type string	pneumonia_resolution_lag_signal int64	earliest_pneumonia_resolution_lag int64	notes string
ABXCS002-TR-0001	S1	0	human_sim	ceftriaxone	0.1	0.9	39.2	39.2	0	90	90	0	30	30	0	0.9	0	vitals_panel	simulated	0	0	baseline
ABXCS002-TR-0002	S1	1	human_sim	ceftriaxone	0.9	0.9	38.7	38.7	0	91	91	0	28	28	0	0.86	0	vitals_panel	simulated	0	0	tracks expected
ABXCS002-TR-0003	S1	2	human_sim	ceftriaxone	0.9	0.9	38.6	38	0.6	90	93	3	29	26	3	0.35	0	vitals_panel	simulated	0	0	first lag signal
ABXCS002-TR-0004	S1	3	human_sim	ceftriaxone	0.9	0.9	38.7	37.6	1.1	89	94	5	30	24	6	0.3	0	vitals_panel	simulated	1	1	onset
ABXCS002-TR-0005	S1	5	human_sim	ceftriaxone	0.9	0.9	38.8	37.2	1.6	88	95	7	32	22	10	0.2	1	vitals_panel	simulated	1	0	later failure
ABXCS002-TR-0006	S2	0	human_sim	azithromycin	0.1	0.8	38.5	38.5	0	92	92	0	26	26	0	0.9	0	vitals_panel	simulated	0	0	baseline
ABXCS002-TR-0007	S2	2	human_sim	azithromycin	0.9	0.8	37.9	37.9	0	94	94	0	22	22	0	0.84	0	vitals_panel	simulated	0	0	response
ABXCS002-TR-0008	S2	4	human_sim	azithromycin	0.9	0.8	37.2	37.2	0	96	96	0	18	18	0	0.83	0	vitals_panel	simulated	0	0	response
ABXCS002-TR-0009	S3	0	human_sim	ceftriaxone	0.9	0.3	39	39	0	90	90	0	30	30	0	0.9	1	vitals_panel	simulated	0	0	severity low
ABXCS002-TR-0010	S4	3	human_sim	ceftriaxone	0.3	0.9	38.7	37.6	1.1	89	94	5	30	24	6	0.3	1	vitals_panel	simulated	0	0	exposure low

ABX-CS-002 Pneumonia Resolution Lag

Purpose

Detect early pneumonia non-improvement when clinical course stops matching the expected timeline under adequate therapy.

Core pattern

severity_index high
abx_exposure_index high
pneumonia_coherence_index drops
spo2_nonresponse_vs_expected stays high
rr_nonresponse_vs_expected stays high
later_icu_escalation_or_treatment_failure_flag appears

Files

data/train.csv
data/test.csv
scorer.py

Schema

Each row is one timepoint in a within series pneumonia course.

Required columns

row_id
series_id
day
host_model
abx_drug
abx_exposure_index
severity_index
temp_c
expected_temp_c
temp_nonresponse_vs_expected
spo2_percent
expected_spo2_percent
spo2_nonresponse_vs_expected
resp_rate_per_min
expected_resp_rate_per_min
rr_nonresponse_vs_expected
pneumonia_coherence_index
later_icu_escalation_or_treatment_failure_flag
assay_method
source_type
pneumonia_resolution_lag_signal
earliest_pneumonia_resolution_lag

Labels

pneumonia_resolution_lag_signal
- 1 for rows at or after first confirmed lag onset
earliest_pneumonia_resolution_lag
- 1 only for the first onset row in that series

Scorer logic in v1

candidate onset point
- abx_exposure_index at least 0.80
- severity_index at least 0.80
- pneumonia_coherence_index at most 0.40
- spo2_nonresponse_vs_expected at least 3.0
- rr_nonresponse_vs_expected at least 3.0
- for two consecutive points
- ignore one point SpO2 artifact then recovery
confirmation
- later_icu_escalation_or_treatment_failure_flag equals 1 later in series

Evaluation

Run

python scorer.py --path data/test.csv

Downloads last month: 8

Number of rows:

Total file size:

16.2 kB