Fetal Cardiotocography Dataset with Expert Consensus Labels

2126 fetal cardiotocograms (CTGs) were automatically processed and diagnostic features measured. The CTGs were classified by three expert obstetricians, with consensus labels assigned for both morphologic patterns and fetal states. The dataset was created by Ayres-de-Campos et al. and is available via OpenML.

Use Cases

• Train a 3-class classifier to predict fetal state (Normal, Suspect, Pathologic) based on CTG features.
• Build a 10-class model to identify specific morphologic patterns (e.g., calm sleep, decelerative pattern) from cardiotocogram data.
• Develop models for automated feature extraction and analysis of short-term and long-term variability metrics from CTG signals.

Strengths

• Contains 2126 records, providing a substantial sample size for model training.
• Features expert consensus labels from three obstetricians, which may improve label reliability.
• Supports both 3-class and 10-class classification tasks, offering flexibility for different experimental designs.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file size are unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Ayres-de-Campos, D., Bernardes, J., Garrido, A., Marques-de-Sa, J., & Pereira-Leite, L.

Collection Method

2126 fetal cardiotocograms were automatically processed and features measured; classifications were assigned by three expert obstetricians.

Time Range

Date of examination is recorded per instance, but overall temporal coverage is unspecified.

Freshness

Last updated date is unknown.

Geography

Spatial coverage is unknown.