MathNet- Matrix returns wrong coefficients to polynomial best fit- C#

Asked Apr 23 '16 at 09:58

Active Apr 23 '16 at 10:04

Viewed 120 times

I used MathNet matrix calculations to calculate the coefficients for a 2nd degree polynomial.

Matrix<double> A = CreateMatrix.DenseOfColumnArrays(xPointsSquared, yPoints, ones);
            Matrix<double> Y = CreateMatrix.DenseOfColumnArrays(yPoints);

Matrix<double> C = A.TransposeThisAndMultiply(A).Cholesky().Solve(A.TransposeThisAndMultiply(Y));

However, the coefficients created by this does not give the accurate curve. It creates a line that nowhere runs along the scatterplot of the data points.

I suspect this is because of the functions using double as opposed to decimal. Can anyone please help me get around this?

edited Apr 23 '16 at 10:04

asked Apr 23 '16 at 09:58

DazedNConfused

I can guarantee this is not related to decimal vs double. – Christoph Rüegg Apr 23 '16 at 10:10
1

Should the `yPoints` in `A` not rather be `xPoints`? – Christoph Rüegg Apr 23 '16 at 10:19
Yes, yes it should. Thanks. Sleep-deprived. – DazedNConfused Apr 23 '16 at 10:40

MathNet- Matrix returns wrong coefficients to polynomial best fit- C#

0 Answers0