which interactions to explore between sex, age, height to predict weight

by Aviad Rozenhek   Last Updated October 19, 2019 20:19 PM - source

As a learning exercise I'm running linear regression to predict person's weight from: sex, age, height. here's a few sample lines

    Sex Age Height  Weight
860 0   50  162.8   58.1
202 1   34  182.3   89.4
580 0   46  171.5   59.9
762 0   40  167.3   55.9
390 1   50  164.8   69.4

having examined the outcome it looks like age is contributing +/- 0.3KG which is very minor. I'm wondering if there is more useful information to be gleaned from age?

  1. I've considered adding interaction variables like sexage or ageheight. should I try other interactions?
  2. I've considered running PCA. would this just rotate my vectors and achieve nothing, or would it add anything?
  3. any other suggestion to try

much obliged



Related Questions