NaN and -infinity when solving system of linear equations

@robs wrote:

Hi there,

I’m trying to do something really simple but am not having any luck. I’m trying to solve a system of linear equations Ax=B

where A, x, and B are 4x4 matrices (x is a rotational and translational transform). I split up B into 4 column vectors and am trying to compute the columns of x separately (I couldn’t figure out how to do all at the same time from looking at the functions available in Math.Net). I’ve tried the following but get NaN and seemingly wrong answers when I compare it to the output from python

        double[][] ddp = new double[][] {
                    new double[] { 38, -144.5, 0, 1 },
                    new double[] { 52, -169.5, 0, 1 },
                    new double[] { 430, -144.5, 0, 1 },
                    new double[] { 66, -144.5, 0, 1 }
        };
        var cp1 = new double[] { 38, 52, 430, 66 };
        double[] q = MathNet.Numerics.LinearRegression.MultipleRegression.QR(ddp, cp1);

I should get an answer like [1, 0, 0, 0] but get [NaN, NaN, -infinity, 0] instead. I’ve also tried using Fit.LinearMultiDimFunc but get an error that the “number of matrix columns must be positive”.

Can someone give me tips on how to proceed?

Thanks in advance!
Robin

Posts: 3

Participants: 2

Read full topic