API reference
np.polynomial (series classes)
Power, Chebyshev, Legendre, Laguerre, Hermite and HermiteE series: classes with domain/window mapping, plus the per-basis module functions.
| Name | Summary |
|---|---|
np.polynomial.Polynomial | Series classes of numpy.polynomial. |
np.polynomial.Polynomial.fit | Class factories. |
np.polynomial.chebyshev | Module functions on raw coefficient arrays. |
np.polynomial.setDefaultPrintstyle | Switches toString() of all series between unicode and ascii output. |
np.polynomial.Polynomial
#new np.polynomial.Polynomial(coef, domain?, window?, symbol = "x") · Chebyshev · Legendre · Laguerre · Hermite · HermiteESeries classes of numpy.polynomial. Coefficients run from lowest degree to highest and are stored as float64 (or complex128). domain is mapped linearly onto window before the series is evaluated. Python operators become methods: add, sub, rsub, mul, truediv, floordiv, mod, divmod, pow, neg, equals. toString() is Python's str() and repr() is repr().
Parameters
coefArrayLike- Series coefficients, lowest degree first.
Returns
series instance
Example
const p = new np.polynomial.Polynomial([1, 2, 3]);
p.call([0, 1, 2]); // => [1, 6, 17]
String(p); // => "1.0 + 2.0·x + 3.0·x²"
p.mul([1, 1]).coef; // => [1, 3, 5, 3]
p.pow(2).coef; // => [1, 4, 10, 12, 9]
p.divmod([1, 1]).map((q) => q.coef.toArray()); // => [[-1, 3], [2]]
p.deriv().coef; // => [2, 6]
p.integ().coef; // => [0, 1, 1, 1]
new np.polynomial.Chebyshev([1, 2, 3]).convert(null, np.polynomial.Polynomial).coef; // => [-2, 2, 6]
new np.polynomial.Polynomial([1, 2]).repr(); // => "Polynomial([1., 2.], domain=[-1., 1.], window=[-1., 1.], symbol='x')"np.polynomial.Polynomial.fit
#Cls.fit(x, y, deg, { domain, window, rcond, w, symbol }) · Cls.fitFull(...) · Cls.fromroots(roots, opts) · Cls.identity(domain?, window?) · Cls.basis(deg, opts) · Cls.cast(series)Class factories. fit computes a least-squares fit, with the domain defaulting to the range of x; .convert() maps the result back to the default domain. p.roots() returns the eigenvalues of the companion matrix, mapped into the domain.
Parameters
xArrayLike- An
NDArray, nested JS array or scalar. yArrayLike- An
NDArray, nested JS array or scalar. degnumber | number[]- Degree, or the list of degrees to include.
Returns
series instance
Example
const f = np.polynomial.Polynomial.fit([0, 1, 2, 3], [1, 3, 5, 7], 1);
f.domain; // => [0, 3]
f.convert().coef; // => [1, 2]
np.polynomial.Polynomial.fromroots([1, 2]).coef; // => [2, -3, 1]
np.polynomial.Polynomial.fromroots([1, 2]).roots(); // => [1, 2]
np.polynomial.Laguerre.basis(2).coef; // => [0, 0, 1]np.polynomial.chebyshev
#np.polynomial.{polynomial, chebyshev, legendre, laguerre, hermite, hermite_e}.<prefix>{add, sub, mul, mulx, div, pow, val, der, int, vander, companion, fromroots, roots, fit, trim, line, domain, zero, one, x} · cheb2poly / poly2cheb ...Module functions on raw coefficient arrays. They keep NumPy's prefixed names (poly, cheb, leg, lag, herm, herme). Only 1-D coefficients are supported. <prefix>int takes { m, k, lbnd, scl }.
Parameters
cArrayLike- 1-D coefficient array.
Returns
NDArray
Example
np.polynomial.chebyshev.chebval(0.5, [1, 2, 3]); // => 0.5
np.polynomial.polynomial.polydiv([1, 2, 3, 4], [1, 2]).map((a) => a.toArray()); // => [[0.75, 0.5, 2], [0.25]]
np.polynomial.legendre.leg2poly([1, 2, 3]); // => [-0.5, 2, 4.5]
np.polynomial.hermite.hermvander([0, 1], 2); // => [[1, 0, -2], [1, 2, 2]]
np.polynomial.polynomial.polyroots([2, -3, 1]); // => [1, 2]TypeScript declaration
np.polynomial.chebyshev: Record<string, unknown>np.polynomial.setDefaultPrintstyle
#np.polynomial.setDefaultPrintstyle("unicode" | "ascii")Switches toString() of all series between unicode and ascii output. p.format("ascii") picks the style for one call.
Parameters
stylestring"unicode"(default) or"ascii".
Returns
void
Example
new np.polynomial.Chebyshev([1, 2, 3]).format("ascii"); // => "1.0 + 2.0 T_1(x) + 3.0 T_2(x)"