linalg

Overview ¶

Package linalg is a require()-able host module providing the vector and matrix primitives that underpin most numeric data-science work: dot products, norms, matrix multiplication, transpose, and the small dense solvers (determinant, inverse, linear solve via Gaussian elimination with partial pivoting) needed for, e.g., closed-form linear regression.

Conventions: a vector is a Lua array of numbers; a matrix is a Lua array of equal-length row arrays. The numeric core operates on []float64 and [][]float64 and is tested directly in linalg_test.go.