[R] Best method

Vladimir Pekar vlapek at vpek.info
Fri Jan 15 11:26:18 CET 2010


I would like to ask for advice about best statistics method for
my problem.
I was done questionnaire about headache.
My data are:
Y - frequency of symptoms occur - times per month in range <0..30>
(where 30 is daily and zero for never)
and independent variables:
X1 - sex - category {M,F}
X2 - age - linear value
X3 - type of work {administrative, manual, ...} 

Which model is "the best" for my problem: to describe weight of
factors in equation "chance_to_present_symptom ~ variables":

Y is not linear value with normal distribution.
Y describe something like chance of to have symptom.
I can't use linear regression, but something like logistic regression. 
Logistic is only for two (or more) Y value which is categorised not

Thank for any advice.


More information about the R-help mailing list