- #1
AlexRaw
- 1
- 0
Hi,
I am working on a data set and I have log-transformed two variables a and b and would like to figure out the relationship between the respective ratios. I have a strong intuition that a/b should have a well-defined range for a given value of lna/lnb, but my maths is poor.
a and be are positive integers with a finite range
0 < a < amax, amax=499639
1 < b < bmax, bmax=837481
I am looking for the range of
a/b
when
ln(a+1)/ln(b+1) = x
0 < x < 1
I know that for x = 1, a/b = 1 and when x = 0, a/b = 0
For anything in between there must be a range.
I'd like to plug in an x and get a range of possible values for a/b.
I have run a regression analysis on my data and figured out the deviations, but I am looking for a cleaner, more general solution that I could potentially also compute when the ranges of a and b are different.
Is this even possible?
Thanks in advance. I'd appreciate any help. :)
I am working on a data set and I have log-transformed two variables a and b and would like to figure out the relationship between the respective ratios. I have a strong intuition that a/b should have a well-defined range for a given value of lna/lnb, but my maths is poor.
a and be are positive integers with a finite range
0 < a < amax, amax=499639
1 < b < bmax, bmax=837481
I am looking for the range of
a/b
when
ln(a+1)/ln(b+1) = x
0 < x < 1
I know that for x = 1, a/b = 1 and when x = 0, a/b = 0
For anything in between there must be a range.
I'd like to plug in an x and get a range of possible values for a/b.
I have run a regression analysis on my data and figured out the deviations, but I am looking for a cleaner, more general solution that I could potentially also compute when the ranges of a and b are different.
Is this even possible?
Thanks in advance. I'd appreciate any help. :)