Khatibi T, Hanifi E, Sepehri MM, et al. - Researchers herein proposed a machine-learning based method that may aid in predicting stillbirth from livebirth and differentiating stillbirth before and during delivery and ranking the features. They proposed a two-step stack ensemble classifier, the first step of which involves classification of the instances into stillbirth and livebirth and the second step involves classification of stillbirth before delivery from stillbirth during the labor. In addition, they proposed a new feature ranking method based on mean reduced accuracy, Gini Index and model coefficients to find high-ranked features. They used the IMAN registry dataset to perform this study and considered all births at or beyond 28th gestational week, including 1,415,623 live birth and 5,502 stillbirth cases. As per experimental results, the proposed stack ensemble classifier outperformed the compared methods with the average accuracy of 90% and AUC of 90%. The two most important features for differentiating livebirth from stillbirth were gestational age and fetal height. In addition, the most important features for classifying stillbirth before and during delivery are hospital, province, delivery main cause, perinatal abnormality, miscarriage number and maternal age.