Compute the arithmetic mean of a one-dimensional single-precision floating-point ndarray, ignoring NaN values and using ordinary recursive summation with extended accumulation. - View it on GitHub