Age: At the core?

Hi all. Experienced data-wrangler who is very new to machine learning. I’m doing everything in R. ** I will update this later with some charts to demonstrate what I mean — Getting the skeleton of the post together. In all models I’ve seen for this, age, or a derived variable, is an important predictor. I spent most of yesterday evening messing around with models to impute age and found some interesting things. First, age’s “ground state” is left-skewed pretty hard, with a very normal looking drop off to the right. There is a bit of a spike of infants. All the models I’ve attempted seem to capture the central tendency pretty well, but the resulting distribution doesn’t seem to capture the skew very well. Additionally, most of the models seem…


Link to Full Article: Age: At the core?