Fitting an Equation to a set of data? Does minitab have the capability?

[JtHall]

Does minitab have the capability to take a formula and fit it to a set of data?

For example I have a data set for force versus time and i have an equation that should fit it, however it has two unknowns. I was hoping i could put this equation into minitab with the data set and have minitab estimate the unkowns by fittign the equation to the data set.

The problem is my experience with minitab is limited and
I dont know how or even if I can do this. Any help would be appreciated.
Thanks

Jon

[Tim Folkerts]

Jon,

Minitab certainly has the ability to fit certain equations to a data set - like straight lines or any number of typical statisitical functions.

If it is something more specific then I don't know that Minitab has any good way of doing that.

On the other hand, there is always Excel. Excel has several built-in fitting possibilities. If that still doesn't cover the function you need, I have used found ways to get Excel to automatically do a least-squares fit worse functions than this using the "solver" ability.

Let me know more specifically what you need, and I'll see what I can think of that might work.


Tim F

[JtHall]

What i start with is a set of 8000+ data points measured by an instrument. Then I have a specific equation that should fit this set of data. It has four unknowns, I can find two by hand. That leaves me with an equation with two unknowns. I need to somehow find these two unknowns. I was hoping I could do so by fitting the equation to my data set.

The solver tool in excel seems promising I have been playing around with it a little but i havent figured it out completely.

Thanks,

Jon

[Tim Folkerts]

What I have done is set up a few extra cells with the coefficients to be adjusted (either the two or all four from your example). Then set up a column beside the data to calcualte the fitted value based on these coefficients and to desired formula. In the next column I calculate the (actual-fit)^2 for each row. Finally, I find the sum of all the values in that last column. As you adjust the coefficients, then the sum changes. When you get the smallest value for the sum, then by definition, you have acheived the least squares fit!

The adjustment can be made by hand - for example, just keep adjusting one and then the other until the sum is as small as you can reasonable get. It helps to make a graph to visualize what is happening. Or the adjustments can be done automatically by "Solver" - although sometimes it "gets of track" and can't automatically narrow in on a solution.


Tim F

[JtHall]

Thanks,

Ill give it a try.

Jon