Multiple Regression with math.net

Question

Hello I am trying to get multiple regression with math.net and I am a little confused.

var xdata = new DenseMatrix(
 new double[,]{{1, 36, 66, 45, 32},
             {1, 37, 68, 12, 2},
             {1, 47, 64, 78, 34},
             {1, 32, 53, 56, 32},
             {1, 1, 101, 24, 90}});

        var ydata = new double[] { 15, 20, 25, 55, 95 };

        var X = DenseMatrix.CreateFromColumns(new[] { new DenseVector(xdata.Length, 1), new DenseVector(xdata) });
        var y = new DenseVector(ydata);

        var p = X.QR().Solve(y);
        var a = p[0];
        var b = p[1];

I guess I don't understand Math.Net, any help with this would be great. Basically I have multiple x and a single y and need to get the coefficient data from them.

What's your target function you want to find the best fitting parameters for? Note that the function has to be linear in these parameters. — Christoph Rüegg, Mar 13 '13 at 20:29
I think least squares would be the best way of doing what I need, Thank you. — user1221399, Mar 13 '13 at 20:38

score 0 · Answer 1 · answered Mar 13 '13 at 21:04

0

From the way you've prepared your matrix (i.e. first column always 1), it seems to me you actually have 4 independent variables and you're looking for a simple regression with just a linear combination of all the independent variables as target function:

y : (x1, ..., x4) -> p0 + p1*x1 + ... + p4*x4

If this is the case, just remove the line var X = ... and instead rename xdata to X, then all 5 parameters will be available in the p vector as expected.

Given the data above, you'll end up with approximately:

y : (x1, ..., x4) -> 123.2 - 8.9*x1 + 2.8*x2 + 3.7*x3 - 4.4*x4

answered Mar 13 '13 at 21:04

Christoph Rüegg

4,626
1
20
34

This line is where I am getting the error: var X = DenseMatrix.CreateFromColumns(new[] { new DenseVector(xdata.Length, 1), new DenseVector(xdata) }); – user1221399 Mar 13 '13 at 21:36
1

Does this solve the problem? Let me know if is still unclear. – Christoph Rüegg Mar 14 '13 at 09:23

Multiple Regression with math.net

1 Answers1

Linked