Based on the data from PALS (Pregnancy and Lifestyle Study), a community-based study of lifestyle on fertility and reproductive outcome from Australia, I built a model of gestational age at birth for non c-section births.

Using 36 features including sex of the fetus, maternal height and weight, maternal education, household demographics, job type, exposure to animals and substances, and reproductive health history, I fit a predictive model to the data using a random forest.

The predicted probabilities given Amy’s specific values are shown in the table and plot below. By contrast to the standard normal distribution that describes the onset of labor in women generally, Amy’s probability distribution is bimodal, with peaks in weeks 38 and 40.

Gestation Week Probability of Birth
37 0.054
38 0.244
39 0.176
40 0.320
41 0.182
42 0.024