Interpolate exponential regression (using Logest function?)

**kriss473** · 03-13-2013, 05:33 PM

Hi,

When I run an exponential regression line through my chart, the following equation is returned: "y=217001e^-.174x"
When I run the logest equation on the same data set, I get the following two numbers: .84039, 217001

What I want to do set up an equation so that I can get the y-values that would result from the regression for given x-inputs. However I am not sure how I would write the equation using the outputs from the logest formula. I know one option would be to manually type in the "y=" equation that I listed in the first line, but I would much prefer to create a dynamic equation that links to the outputs of the logest function so that I don't have to manually type this every time I replicate the analysis.

Any ideas?

**MrShorty** · 03-13-2013, 05:48 PM

According to the help file http://office.microsoft.com/en-us/ma...829.aspx?CTT=1 the LOGEST function is y=bm^x, where the chart trendline uses the equation y=be^(ax).

Comparing the two equations with a little algebra chart trendline y=b(e^a)^x, logest y=b(m)^x and you should see that m from the logest function is equal to e^a from the chart trendline (which, if you put e^-.174 into a calculator or spreadsheet cell, you will get 0.84030 -- the same as logest to within 3 significant figures).

So, the two equations should be equivalent, if you use the correct form for each equation.