XTRACTIS FOR PREDICTIVE MEDECINE
CARDIOTOCOGRAPHIC IDENTIFICATION OF FETAL HEART PATHOLOGIES
BENCHMARK: XTRACTIS VS. RANDOM FORESTS, BOOSTED TREES & NEURAL NETWORKS
This study aims to solve the following problem:
How to make an automated medical diagnosis of fetal heart disease from signal characteristics of fetal heart rate and uterine contractions?
GOALS
- Decrease prenatal mortality.
- Help the medical profession make an earlier and more individualized decision, thanks to rapid and systematic diagnoses.
- Avoid possible neurological sequelae for the fetus.
XTRACTIS RESULTS
A predictive model that is:
- Intelligible. 56 gradual unchained rules, based on 18 predictors.
- Robust. Excellent predictive performance.
- Efficient. Instant prediction (online or offline).
CONTENTS
- Problem Definition
- Xtractis Solution
- Top-Model Induction
- Prediction for 1 Unknown Case of Testing
- Top-IVE Benchmark
