Adapting LINEST Muliti-variable regression - change one variable to be Logarithmic

[GMcCabe]

I currently have a multi-variable linear regression model that I am using with the LINEST excel function. The issue I have is that one of the variables is skewing the data too much and I am trying to find away to adapt just that variable to be logarithmic instead of linear while keeping the remaining variables linear. Is there an easy way to do this?

My initial thoughts were to input a dummy outcome for that variable and then have a second Logarithmic model that would be used for just that one variable, but I'm trying to find an easier way around this.

Any ideas would be greatly appreciated!!!

[shg]

=linest(yyy, choose({1,2,3}, xxx1, xxx2, ln(xxx3)))

[GMcCabe]

So you're saying all that I would need to do is add LN to the variable I would want to behave logarithmic-ally?