Impact of Cross-Validation on Machine Learning Models for Early Detection of Intrauterine Fetal Demise