multiple variables regression

[plumberef]

I have a data set of 4 * 10.

I would like to establish a relationship y = f(X1,X2,X3) - linear, power etc etc with the respective R2.

Are there macros available for this ?

Many thanks.

[MrShorty]

I don't think I've seen generic, robust "macros" for this. Most macros for this are specific to a project or spreadsheet.

If the relationship is "linear" [in the linear algebra sense, meaning v(y)=a1*f1(x1)+a2*f2(x2)+a3*f3(x3)+... kind of relationships], then you can simply use the LINEST() function to do a basic least squares regression http://office.microsoft.com/en-us/ex...005209155.aspx No macros necessary for the basic regression.

If there are any "non-linear" parts of the function, or you want an objective function (OF) other than a least squares OF, you will need to use a numerical algorithm like Solver to obtain the coefficients. Again, no macros necessary for the basic regression.

If you need more help, we will need a more specific example of what you are doing.

[plumberef]

I don't think I've seen generic, robust "macros" for this. Most macros for this are specific to a project or spreadsheet.

If the relationship is "linear" [in the linear algebra sense, meaning v(y)=a1*f1(x1)+a2*f2(x2)+a3*f3(x3)+... kind of relationships], then you can simply use the LINEST() function to do a basic least squares regression http://office.microsoft.com/en-us/ex...005209155.aspx No macros necessary for the basic regression.

If there are any "non-linear" parts of the function, or you want an objective function (OF) other than a least squares OF, you will need to use a numerical algorithm like Solver to obtain the coefficients. Again, no macros necessary for the basic regression.

If you need more help, we will need a more specific example of what you are doing.

Mr Shorty,

Noted and thanks.

I am studying some companies' annual reports and would like to find a relationship between Free Cash Flow and parameters like sales, earnings, return on equity, Price/Earning, Price/Sales etc.

Cheerio.