Cervical Dilation Classification from Electrohysterography and Clinical Features

648 segments of 120-second electrohysterography recordings from 72 low-risk labors, labeled for cervical dilation stage. The dataset, published by figshare admin karger on 2026-04-28, was used to develop a machine learning model achieving perfect classification on an independent test set. It combines 21 EHG signal descriptors with four clinical variables.

Use Cases

• Classifying cervical dilation stage (Low, Moderate, Advanced) based on electrohysterography signal features.
• Evaluating feature importance for labor progression using clinical variables like maternal age and gestational age.
• Training ensemble classifiers, such as Bagged Trees, for time-series physiological signal analysis.
• Optimizing feature subsets for medical diagnostics using genetic algorithm approaches.

Strengths

• Includes 648 data segments derived from 72 patient recordings, providing a basis for model development.
• Model performance reported with an F1 score of 1.000 on an independent test set of 194 segments.
• Combines 21 objective EHG signal descriptors with 4 clinical variables, offering a multi-modal feature set.

Limitations

• Dataset is very small at 40.1 KB, indicating limited scope and sample size.
• Row count and column-level documentation are unknown, requiring inspection after download.
• Data is from a specific pilot study of low-risk labors, which may limit generalizability.

Provenance

Source

figshare

Collection Method

Analysis of single-channel EHG recordings from low-risk labors, segmented and labeled by cervical dilation stage.

Freshness

Last updated 2026-04-28 07:55:27.