Hi,
I came across an interesting study that did the following:
'number of births in the last 5 years and the number of household members were the “most important” features for predicting whether a mother reported the death of a neonate. Out of the 20 models (10 countries and 2 DHS surveys per country) trained to predict neonatal mortality, the number of births in the last 5 years ranked first in all models except Burkina Faso 2003, Tanzania 2015, and Zambia 2007, where the feature ranked second.'
I am wondering whether this is possible to do using Stata.
Please let me know if you are aware of any process similar to this.
Thank you.
0 Response to Ranking the most important predictor variables from a series of regression models.
Post a Comment