Fetal Health Classification.pdf

Cardiotocography (CTG) is used to monitor fetal heart rate, ensuring fetal well-being during high-risk pregnancies. Machine learning and deep learning techniques can automate this task, reducing diagnostic errors.

Machine Learning-Based Approaches

Random Forest Classifier: ensemble of decision trees, achieving 99.02% accuracy in classifying CTG data as healthy or unhealthy.
Other algorithms used: Support Vector Machine (SVM), Artificial Neural Network (ANN), Classification and Regression Trees (CART), K-Nearest Neighbors (K-NN), Logistic Regression, C4.5, and Reduced Error Pruning Tree (REP Tree).
Techniques: bagging ensemble classifier, bootstrap aggregation, and 10-fold cross-validation.

Deep Learning Techniques

Artificial Neural Network (ANN): used to categorize CTG data with simple logistics.
Multimodal Convolutional Neural Network: used on data from 35,429 births to detect life-saving information from CTG data.
Deep learning algorithms: used to analyze CTG data, recognizing outliers, executing classification algorithms, and clustering.

Key Findings

Random Forest Classifier and ANN showed exceptionally good results in classifying CTG data.
Bagging approach with Random Forest showed good performance in classifying fetal health.
* Deep learning techniques can be used to detect chromosomal abnormalities and potential risks of euploidy.

Conclusion

Machine learning and deep learning techniques can be used to classify fetal health from CTG data, reducing diagnostic errors and improving prenatal care. Random Forest Classifier and ANN are promising algorithms for this task, while deep learning techniques can be used to detect life-saving information from CTG data.