def r_squared(returns, benchmark):
""" measures the straight line fit of the equity curve """
# slope, intercept, r_val, p_val, std_err = _linregress(
_, _, r_val, _, _ = _linregress(
_utils._prepare_returns(returns),
_utils._prepare_benchmark(benchmark, returns.index))
return r_val**2
def polynomial_fit(x, y, n):
"""
Returns a polynomial fit for y given x of order n
with an R-squared score of the fit
Parameters
-----------
x : numpy array
x data for polynomial fit.
y : numpy array
y data for polynomial fit.
n : int
order of the polynomial fit.
Returns
----------
polynomial_coefficients : numpy polynomial
List of polynomial coefficients
R2 : float
Polynomical fit coeffcient of determination
"""
try:
x = np.array(x)
except:
pass
try:
y = np.array(y)
except:
pass
assert isinstance(x, np.ndarray), 'x must be of type np.ndarray'
assert isinstance(y, np.ndarray), 'y must be of type np.ndarray'
assert isinstance(n, int), 'n must be of type int'
# Get coeffcients of polynomial of order n
polynomial_coefficients = np.poly1d(np.polyfit(x, y, n))
# Calculate the coeffcient of determination
slope, intercept, r_value, p_value, std_err = _linregress(y, polynomial_coefficients(x))
R2 = r_value**2
return polynomial_coefficients, R2
