In my dataset, there are some individual features such as height, weight, and BMI, as well as some measured features such as leg length or thigh circumference. I want to modify my measured properties based on for example BMI (combined weight and height). For example, the X size for leg length may be normal for a height of 180 but not normal for a height of 160. I thought of calculating the average total BMI and then doing this for each measured feature:
(feature measured for the person i * mean BMI)/BMI of person i
it's correct? Or is there a feature engineering method for this? Thanks.