0
$\begingroup$

I'm reading a scientific paper and an equation of the following form appears:

x = y log (z).

I know what y and z are in my own data set. How do I solve for x?

I'm used to logarithms of the form log28, and then I can work out that the base is 2 and hence the answer is 3. How can I go about working out this unfamiliar (to me) form of logarithm? What's the base?

Edit: May as well provide a link to the paper. It's here, and the relevant part is on p264.

$\endgroup$
  • 3
    $\begingroup$ An unspecified base is 10. so it would be $y log_{10}(z)$ $\endgroup$ – Stretch Maniac Nov 11 '14 at 1:49
  • 4
    $\begingroup$ An unspecified base is also quite often $e$. $\endgroup$ – Hugh Denoncourt Nov 11 '14 at 1:50
  • $\begingroup$ That would be denoted $ln$ wouldn't it? $\endgroup$ – Stretch Maniac Nov 11 '14 at 1:51
  • 3
    $\begingroup$ In an undergraduate American algebra textbook, yes. In a scientific or mathematical paper, unspecified log can be $e$ or even $2$, even though $\ln$ and $\text{lg}$ are standard elsewhere. $\endgroup$ – Hugh Denoncourt Nov 11 '14 at 1:52
  • 1
    $\begingroup$ Since it's log-odds, the best guess seems to be $e$. It also looks like there is a parameter scaling the log-odds so that it doesn't matter what base is used. The wiki en.wikipedia.org/wiki/Logit supports multiple possibilities (including 2). stats.gla.ac.uk/glossary/?q=node/272 supports $e$ as standard for log-odds. Also, with $e$ as base, log-odds is the inverse of the logistic function. $\endgroup$ – Hugh Denoncourt Nov 11 '14 at 3:09
2
$\begingroup$

The paper doesn't explicitly mention a base, but later on there's a reference to exponentiation, which indicates that the base is e.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.