How to Expand Log(a+b) for Best Fit Analysis?

[Saoist]

anyoen know how to expand this? i can't think of any obvious way...

 

[mathman]

What kind of result are you looking for - functions of a and b separately? As it stands, it is as simple as possible.

 

[CRGreathouse]

There's not much you can do. In some cases, it's useful to factor it as \log a+\log(b+1), but in general there's nothing simpler than the way you wrote it.

 

[Saoist]

i have a deceptively simple question you see:

X^3 = (cY+d)^2

where c and d are constants, with x and y the variables. how would you plot the 2 variables as a straight line graph. I'm having an idiocy attack and can only think "log it..."

 

[Learning Curve]

Take the log of Y and graph x, log y.

 

[Saoist]

that doesn't plot that relationship as a straight line though does it?

i was under impression you had to transform [said equation] into a y=mx+c type form

 

[NateTG]

that doesn't plot that relationship as a straight line though does it?
i was under impression you had to transform [said equation] into a y=mx+c type form

You can't plot things like x^3=y^2 as a straight line on a normal graph.

 

[Learning Curve]

I didn't mean that would give you a formula, but if you had a set of data, you could find the regression by plotting x, log y. It's not the answer but it's a way to get it.

 

[uart]

that doesn't plot that relationship as a straight line though does it?
i was under impression you had to transform [said equation] into a y=mx+c type form

No, none of log-log, log-linear or linear-log will make that equation a straight line.

What's the full context of the problem, do you have a number (more than 2) of x,y points and you wish to find constants c and d that give the "best fit" in some particular sense?