Probability of an outcome given multiple attributes

[redeemer90]

Hi,

Lets say I have the following attributes...
Body fat percentage, Body mass index
Outcome = obese (yes / no)
I have data for all attributes and outcomes. What I'd like to do is a joint probability. For example I'd like to know what the probability is of someone being obese given their BMI and body fat percentage.

I can calculate the correlation between the attributes and the outcomes. My intuition tells me that I need to combine the correlation in some way, but can't seem to find any information on this.

Can someone point me in the right direction?

- Sid

 

[moonman239]

I don't know about joint probability, but I'd say something like this would do:
P(being obese) = P(person with BMI x being obese) * P(person with Body Fat Percentage n being obese)

http://en.wikipedia.org/wiki/Joint_probability_distribution

 

[EnumaElish]

You should look into regression analysis with a limited (binary) dependent variable, "limdep."

See "logit" and "probit" under: http://en.wikipedia.org/wiki/Limited_dependent_variable

 

[moonman239]

It sounds like you're actually looking at discrete choice modeling: http://en.wikipedia.org/wiki/Discrete_choice

Found this under the article that Enuma linked to.

 

[redeemer90]

Thanks for all the help. Will look into all the links now and see which one suits my needs.

- Sid