Mark Rogoyski

@Beakerboy Thanks for sharing your work. Good stuff. What do you think about putting the unit circle in a new namespace like Trigonometry or something?

My preference would be that eigenvalues and eivenvectors are separate functions that return separate things. Eigenvalues can lazy load the values. That way, when eigenvectors is called, if eigenvalues has...

I imagine there will end up being multiple NxN algorithms implemented, perhaps with an appropriate method being chosen automatically based on the characteristics of the matrix if the user does...

Any thoughts on what the default return value should be from ```Matrix->eigenvalues()```? Two options spring to mind: * DiagonalMatrix with eigenvalues along the diagonal (Currently doing this in Eigenvalue class)...

If you send a one-dimensional array of values to MatrixFactory::create you will get a DiagonalMatrix, so that functionality already exists.

For LU Decomposition, the return value is array containing [L, U, P, A]. I don't know why I put A in there, but L U P is what I think...

Hi @andrewdalpino, Yes, the QR decomposition is on the roadmap. I will prioritize it since multiple people have now requested this feature. Thanks for your interest in MathPHP and providing...

Hi, Thank you for your continued interest in MathPHP. I apologize that we are lacking some functionality you'd like to use. I am not familiar with how to compute the...

Hi, thanks for your interest in MathPHP. Yes, the intention is to support all the useful statistical regression methods. Since there is now a feature request for these we can...

@sbrbot Can you explain what is the optimization you think the MathPHP inverse is not using? The basic strategy is to reduce to RREF and then find the inverse. We...