[R] Power law, lognormal end exponenntial statistical testing in one sample population
ΠΑΝΤΟΠΟΥΛΟΣ ΓΕΩΡΓΙΟ
ΠΑΝΤΟΠΟΥΛΟΣ ΓΕΩΡΓΙΟ
Mon Feb 11 14:45:38 CET 2008
Hello,
My name is George Pantopoulos, i am a phd student in the Dept. of Geology,
University of Patras, Greece.
I am studying the statistical behaviour of bed thickness datasets taken
from outcrops. Until now, 4 statistical distributions seems to fit my
datasets: power law, lognormal, lognormal mixture with 2 modes and
exponential.
I already used the MIX package of R to detect a possible 2 mode lognormal
mixture in the datasets.
My questions are :
Can i do ONE POPULATION non-parametric tests (ks or x2) in R for the above
distributions, with the distribution parameters NOT estimated from the
data ? (especially for the ks test). If yes , is it possible for someone
to show me the exact steps that i must follow in R ? Also, must i generate
artificial populations of data to do what i want ?
The datasets are in text files, one variable (bed thicknesses only).
Please if somebody knows how to do something of the above in R , it would
be valuable for my work.
Thank you.
More information about the R-help
mailing list