Scipy Integrate Function which takes two functions as arguments that depend on the same variable

Vote count:

0

I want to numerically integrate the following function:

I know the function values of f over a grid called grid_y and grid_posterior. To evaluate the entire function, I use linear interpolation as follows:

Af = lambda x: scipy.interpolate.interpn(tuple([grid_y,grid_posterior]),f,x)

The function G(y,z) is written as:

lambda z: y**alpha * np.exp(mu) * z. 

H(y,\mu,z) is called:

updated_posterior =  lambda z: ( posterior**np.exp(-np.log(z)/sigma**2/
( np.sqrt(2*np.pi)*sigma*y*np.exp(mu)*z ) ). 

To evaluate the integral, I create the function:

integrand = lambda z: Af( np.array([y**alpha * np.exp(mu[0]) * z, 
                         updated_posterior(z)])) * phi.pdf(z)

and then pass this function to the scipy.integrate.fixed_quad module:

 fixed_quad(integrand, a, b)

My procedure unfortunately results in an error. The fixed_quad function returns the following text:

*** ValueError: The requested sample points xi have dimension 5, but this
    RegularGridInterpolator has dimension 2

Does somebody know the source of the error or has a better suggestion in how to evaluate the integral?

asked 36 secs ago

fabian

94

Scipy Integrate Function which takes two functions as arguments that depend on the same variable