Aug. 26th, 2005

gusl: (Default)
I'm very impressed with the common-sense idea of testing your blood in order to create a personalized diet.

But unfortunately, Dr Greg Tefft's sites (1 2 3) have an unhealthily high level of BS.

It's not even clear what the "Total Body Chemistry Test kit" is! Is it a DIY pee-in-a-cup chemistry set?

In any case, this got me interested in what machine learning can offer to the problem of inferring which diets lead to which diseases. Clinical data is probably far from being standardized, with all kinds of hidden biases and incomplete data. But even then, computers should be better than humans at analyzing this data, and machine learning has already explicitly dealt with several of these problems (e.g. learning from unlabelled data)! Machine Learning experts are the people who are good at induction, because their methods are always getting evaluated and improved (unlike statisticians, AFAIK)!
So unlike Dr. Tefft, I'm skeptical of this "best scientific knowledge".

Do nutritionists offer similar tests?

This is an area I'd very much like to work on.

February 2020

S M T W T F S
      1
2345678
9101112131415
16171819202122
23242526272829

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags