Hi all,

I wonder if someone could point me in the right direction here...I have done a regression analysis where only the x-axis was logged (log10) to achieve the best straight line. So the equation I have looks like...

y = b*log(x)+a

I'm stuck now with how to antilog this? If its even possible? If both axis were logged I'd have no problem, but have just been stumped by this.

Many thanks for any help!

Is everyone else as unsure as me then?

why do you need antilog here?

Richie is right, you are fine, if this transformation on one axis only helped you, i mean produced a model that satisfies the assumptions just interpret it the usual way, but instead of x you will have log(x).

Thanks for the responses...my desire to antilog it is to enable its comparison to previously published data.

:confused: