Cardiotocographic Identification of Fetal Heart Conditions

Benchmark vs. Logistic Regression, Random Forests, Boosted Trees & Neural Networks

How to make an automated —yet transparent— medical diagnosis of fetal heart condition from signal characteristics of fetal heart rate and uterine contractions?

Goals & benefits

Identify parameters requiring increased vigilance and improve medical knowledge by helping cardiologists understand the causal relationships between specific cardiotocographic features, their combination, and the presence of an abnormality.

Help the medical profession to make earlier and more personalized decisions by means of rapid, systematic, and explainable diagnoses. Extend access to high-level diagnoses even in medical deserts

Decrease prenatal mortality and avoid possible neurological sequelae for the fetus.


We get a Predictive Model that is:


A Decision System composed of 56 unchained gradual rules, each using some of the 18 variables that XTRACTIS identified as significant, including weak signals (out of the 21 Potential Predictors characterizing the fetal cardiotocograms and uterine contraction signals).


Good Real Performance on External Test.

Efficient & Operational.

Running in real-time up to 70,000 predictions per second (i7 @2.5GHz with 8 physical cores), offline or online (API).

Use Case 09/2022 (v2.3)

Results by
XTRACTIS® GENERATE 12.1.41978 (05/2022)


  1. Problem Definition
  2. Xtractis Solution
  3. Top-Model Induction
  4. Explained Predictions for 4 cases
  5. Top-Models Benchmark