Non-negative Matrix Factorization (with nalgebra)

Takes a matrix V of size R*C, splits it into two smaller matricies W (R*K) and H (K*C) such that W * V approximates V.

• Uses multiplicative update rule
• Generic over allocation strategy, matrix size, and scalar type.
• Doesn't do allocation in the loop :)
• 195 lines of code!
• PRs welcome!