How to find when function start to behave as 1/(powerx,2)

Hi .
I have some measured data (see attached files ) i need to find when it function start to behae as 1/(x^2).
is any one can assist me please ?
funkzia.PNG
data.xlsx

Welcome to the forum!

Will you please attach a sample Excel workbook? We are not able to work with or manipulate a picture of one and nobody wants to have to recreate your data from scratch.

1. Make sure that your sample data are REPRESENTATIVE of your real data. The use of unrepresentative data is very frustrating and can lead to long delays in reaching a solution.

2. Make sure that your desired solution is also shown (mock up the results manually).

3. Make sure that all confidential data is removed or replaced with dummy data first (e.g. names, addresses, E-mails, etc.).

4. Try to avoid using merged cells as they cause lots of problems.

Unfortunately the attachment icon doesn't work at the moment, so to attach an Excel file you have to do the following: just before posting, scroll down to Go Advanced and then scroll down to Manage Attachments. Now follow the instructions at the top of that screen.

Please pay particular attention to point 2 (above): without an idea of your intended outcomes, it is often very difficult to offer appropriate advice.

comment removed by GK

Your sample file only contains your raw data and a chart. There are no attempts at any kind of analysis, and no description of how you would like to perform this analysis.

It seems like the first step would be to add a column =A2^-2 (copied down) and add that to your chart. This will give you something to compare to.
My first observation, with that column added, is that the function at least needs a scaling factor (=$F$1*A2^-2 where F1 contains the scaling factor). I manually tried a few different values and settled on 1E9, which was close enough to visually compare in the chart the two curves.

For a more analytic algorithm, I could see adding an "error" column [((yest-ymeas)/ymeas)^2 or ln(yest/ymeas)^2 or whatever makes sense to the problem at hand) and using that column to develop a more analytic measure of "when does measured y begin to look like y=A*X^-2".

I could also see getting even more statistically involved by using regression techniques (LINEST()) perform a regression, then doing an "outliers" analysis and using that to determine which points are outliers. Based on where the outliers tend to congregate, you should be able to see which regions do and do not "fit" the y=a*x^-2 equation.

That's my overview of the problem. Where do you want to go with this, and what do you need help with?

Can you please explane about "regretion techniques"?

Depends on how much of an explanation you need.

Basically, regression techniques are more standardized, formal expressions of the basic "manually try different scaling factors until you get a curve that matches" process that you just did. If you are completely unfamiliar with regression, I would suggest that you start with an introduction to statistics text/course/page, as basic linear regression is a common part of introduction to statistics. Something like this: http://stattrek.com/regression/linea...?Tutorial=Stat
or http://onlinestatbook.com/2/regression/intro.html
or http://people.duke.edu/~rnau/regintro.htm
or https://www.youtube.com/watch?v=zPG4NjIkCjc

From there, you can expand the concept to other "linear equations", where linear is not longer about "straight lines" (like y=A*x^-2). A full course on regression (or two or three) is probably more than can be included in an internet forum post, so I will leave you to research the techniques elsewhere. As you come across different techniques and want to know how to program those into the spreadsheet, let us know and we can help with that.