Datasets:

ClarusC64
/

long-covid-closed-loop-recovery-control-v1.0

scenario_id string	control_sequence_alignment_score float64	control_horizon int64	feedback_response_score float64	intervention_timing_score float64	adaptation_latency float64	control_stability_margin float64	sequence_divergence_margin float64	controller_confidence float64	recovery_consistency_score float64	control_recalibration_count int64	terminal_pathway_state string	successful_recovery_control int64
LC001	0.84	6	0.81	0.79	0.18	0.77	0.22	0.82	0.86	1	stable_recovery	1
LC002	0.79	5	0.76	0.74	0.22	0.72	0.28	0.78	0.8	2	stable_recovery	1
LC003	0.31	4	0.38	0.42	0.71	0.26	0.74	0.35	0.33	5	relapsing	0
LC004	0.27	3	0.32	0.36	0.78	0.21	0.81	0.3	0.28	6	deteriorating	0
LC005	0.73	6	0.7	0.68	0.27	0.66	0.34	0.74	0.75	2	stable_recovery	1
LC006	0.44	5	0.49	0.46	0.59	0.38	0.62	0.42	0.45	4	persistent	0
LC007	0.88	7	0.84	0.82	0.15	0.8	0.18	0.86	0.89	1	stable_recovery	1
LC008	0.36	4	0.41	0.39	0.68	0.3	0.7	0.37	0.34	5	relapsing	0
LC009	0.69	5	0.72	0.71	0.31	0.64	0.36	0.7	0.73	2	stable_recovery	1
LC010	0.52	5	0.55	0.5	0.48	0.46	0.51	0.5	0.49	3	persistent	0
LC011	0.81	6	0.78	0.76	0.2	0.74	0.25	0.8	0.83	1	stable_recovery	1
LC012	0.29	3	0.35	0.33	0.75	0.24	0.78	0.32	0.3	6	deteriorating	0
LC013	0.76	6	0.73	0.7	0.24	0.69	0.31	0.75	0.78	2	stable_recovery	1
LC014	0.41	4	0.44	0.47	0.63	0.36	0.65	0.4	0.39	4	relapsing	0
LC015	0.86	7	0.83	0.8	0.17	0.79	0.2	0.84	0.87	1	stable_recovery	1
LC016	0.33	4	0.37	0.35	0.72	0.28	0.76	0.34	0.31	5	deteriorating	0
LC017	0.71	5	0.69	0.67	0.29	0.65	0.35	0.72	0.74	2	stable_recovery	1
LC018	0.47	5	0.5	0.48	0.55	0.41	0.58	0.45	0.43	4	persistent	0
LC019	0.83	6	0.8	0.78	0.19	0.76	0.23	0.81	0.84	1	stable_recovery	1
LC020	0.38	4	0.42	0.4	0.66	0.32	0.68	0.39	0.36	5	relapsing	0

What this dataset does

This dataset tests whether a model can identify successful closed-loop recovery control in a synthetic Long Covid recovery setting.

The task is not diagnosis.

The task is control success prediction.

Core stability idea

A recovery plan may begin well but fail if feedback is poor, timing is wrong, adaptation is slow, or the control sequence diverges.

This dataset tests whether models can identify when a recovery control loop is likely to stabilize the system.

Prediction target

The target column is:

successful_recovery_control

Labels:

0 = failed control
1 = successful control
Row structure

Each row represents a synthetic Long Covid recovery control state.

Columns:

scenario_id
control_sequence_alignment_score
control_horizon
feedback_response_score
intervention_timing_score
adaptation_latency
control_stability_margin
sequence_divergence_margin
controller_confidence
recovery_consistency_score
control_recalibration_count
terminal_pathway_state
successful_recovery_control
Files
data/train.csv
data/test.csv
scorer.py
README.md
Evaluation

Predictions should use this format:

scenario_id,prediction
LC101,1
LC102,0

Run:

python scorer.py predictions.csv data/test.csv

The scorer reports:

accuracy
precision
recall_correct_control_success
f1
false_effective_control_rate
confusion_matrix

Primary metric:

recall_correct_control_success

Secondary metric:

false_effective_control_rate
Structural Note

This dataset is part of the Clarus / SIOS synthetic benchmark series.

It extends intervention competition into closed-loop recovery control.

The benchmark evaluates whether models can distinguish a promising intervention from a stabilizing control sequence.

Production Deployment

This dataset is synthetic.

It should not be used for clinical decision-making.

A production version would require longitudinal patient-level data, intervention sequences, feedback signals, and independently validated recovery outcomes.

Enterprise & Research Collaboration

Future versions may incorporate:

patient-level recovery time series
intervention timing
relapse events
autonomic feedback
immune profiling
metabolomics
symptom trajectories
control adaptation logs
treatment response data
License

MIT

Downloads last month: 16

Number of rows:

Total file size:

9.71 kB