go-approx
You are reading the documentation of an outdated version, whose text may no longer be accurate. Go to the latest version.
go-approx is the home of approximation in the parity plan: interpolation, splines, polynomials, and later Chebyshev and AAA approximation and scattered data.
| Package | Status | Repository |
|---|---|---|
| interp | released v0.2.0 | go-approx/interp |
| poly | released v0.1.0 | go-approx/poly |
What is verified
From STATUS.md:
numpy.interp, interp1d, PCHIP, Akima and poly1d are bit-identical to
NumPy and SciPy; splines agree within 1e-15 to 1e-10. interp v0.2.0 adds Splrep
(FITPACK’s legacy splrep) and is faster than SciPy on every case measured
except two still open: RegularGridInterpolator linear evaluation at 10^3 to
10^4 points (0.95 to 0.98x) and shuffled-query cubic evaluation at 10^4 points
(0.94x). The speed comparison with the Python reference is in
interp’s docs/perf.md.
Where to read more
- Landing: go-approx.github.io
- Documentation: go-approx.github.io/docs
- Organisation: github.com/go-approx
Created on by tannevaled
Modified on by tannevaled