Algorithms Non-negative least squares
this algorithm takes finite number of steps reach solution , smoothly improves candidate solution goes (so can find approximate solutions when cut off @ reasonable number of iterations), slow in practice, owing largely computation of pseudoinverse ((aᴾ)ᵀ aᴾ)⁻¹. variants of algorithm available in matlab routine lsqnonneg , in scipy optimize.nnls.
many improved algorithms have been suggested since 1974. fast nnls (fnnls) optimized version of lawson—hanson algorithm. other algorithms include variants of landweber s gradient descent method , coordinate-wise optimization based on quadratic programming problem above.
Comments
Post a Comment