There are a few important variables in my model with high variability across biological repeats. Is there any way to deal with this situation?

Let's take blood pressure as an example. I measure the blood pressure of one patient 2 times. Because of the limit of my equipment, I may get different numbers for the same patient. Moreover, I don't know the exact systematic errors for my measurements.

How to deal with this kind of features in building predictive models? If a new patient comes in, I can only measure his blood pressure once, how should I use this feature in my model to give a stable prediction?

  • $\begingroup$ Can you model it with high standard deviation? $\endgroup$ – SmallChess Mar 11 '17 at 9:46
  • $\begingroup$ How to? For example, how to let SVM know this feature has high standard deviation? $\endgroup$ – user2149631 Mar 12 '17 at 22:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.