How to use SwarmPackagePy in regression in python code of sklearn?

Question

I'm able to draw a 3D animation of Firefly algorithm using SwarmPackagePy library.

I want to use this algorithm to optimize the hyperparameter in Gaussian Process Regression(GPR). For this, I defined the optimizer of GPR as:

alh = SwarmPackagePy.fa(50, tf.easom_function, 0, 16, 2, 10, 1, 1, 1, 0.1, 0, 0.1)

animation3D(alh.get_agents(),tf.easom_function, 10,-10)

Then I used this optimizer (alh) in GPR as follows:

gp = GaussianProcessRegressor(kernel=kernel, alpha=1.5, optimizer=alh, n_restarts_optimizer=5)

However, after running the python code, I get an error as follows:

ValueError: Unknown optimizer <SwarmPackagePy.fa.fa object at 0x0982A3B0>.

Am I doing the wrong way? What could be the cause of the error?

Thank you!

Answer 1

As the documentation of sklearn says, optimizer parameter expects a callable. However, SwarmPackagePy.fa is not a callable. Because it is neither a method, nor a class that implements __call__ method as can be seen from here:

https://github.com/SISDevelop/SwarmPackagePy/blob/master/SwarmPackagePy/fa.py

You can write your own __call__ method in that file, by using the same signature as:

def __call__(obj_func, initial_theta, bounds):
# you need to write
# * 'obj_func' is the objective function to be maximized, which
#   takes the hyperparameters theta as parameter and an
#   optional flag eval_gradient, which determines if the
#   gradient is returned additionally to the function value
# * 'initial_theta': the initial value for theta, which can be
#   used by local optimizers
# * 'bounds': the bounds on the values of theta
....
# Returned are the best found hyperparameters theta and
# the corresponding value of the target function.
    return theta_opt, func_min