Fetal Cardiotocography Dataset with Expert Consensus Labels for 2,126 Examinations

2,126 fetal cardiotocograms (CTGs) were automatically processed and their diagnostic features measured. The CTGs were classified by three expert obstetricians, with consensus labels assigned for both morphologic patterns (A, B, C, etc.) and fetal states (Normal, Suspect, Pathologic). The dataset was created by Ayres-de-Campos et al. and is available via OpenML.

Use Cases

• Classifying fetal state (Normal, Suspect, Pathologic) based on features like baseline value, accelerations, and decelerations.
• Predicting morphologic pattern (e.g., calm sleep, decelerative pattern) from histogram statistics and variability metrics.
• Developing multi-class (10-class) models for detailed CTG pattern recognition using the CLASS column.
• Analyzing relationships between uterine contractions, fetal movement, and expert-assigned diagnostic labels.

Strengths

• Contains 2,126 real-world fetal cardiotocogram records.
• Features expert consensus labels from three obstetricians for two classification schemes.
• Includes over 20 measured diagnostic features such as baseline values, accelerations, decelerations, and histogram statistics.

Limitations

• Row and column counts are unspecified, limiting suitability assessment.
• Last update date is unknown; freshness unverified.
• Description metadata is limited; actual data quality and completeness require manual inspection after download.

Provenance

Source

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

Collection Method

Fetal cardiotocograms were automatically processed, and features were measured. Three expert obstetricians classified each CTG, with a consensus label assigned.

Time Range

null

Freshness

null

Geography

null