Issue
I am writing some code to fit distributions to data and since for the Pareto 1 distribution, theta was fixed, I wrote the following code:
class ParetoI:
alpha = None
theta = None
def __init__(self, alpha, theta=.5):
self.alpha = alpha
self.theta = theta
return
def pdf(self, x):
return ParetoI.pdf(self.alpha, x, self.theta)
# Some other code
@staticmethod
def pdf(alpha, x, theta=.5):
return alpha * theta ** alpha / x ** (alpha + 1)
Then in my main routine, after I have fit a distribution to the given sample data, I try to plot the density:
# par1 = ParetoI(some_args)
x_range = np.linspace(.5, 80, 200)
plt.plot(x_range, par1.pdf(x_range), label='Pareto I')
But I get the following error:
TypeError: pdf() missing 1 required positional argument: 'x'
I assume that this is somehow because I call the generalized, static function from the par1.pdf. How can I fix this code? Am I simply not allowed to have a static function with the same name as a class method? I guess this is easily fixable by simply removing the static functions, but I thought it might be handy to be able to use these distributions without instantiating an object for further use. Is this bad design?
Solution
So, as suggested above, the Python way to write that is:
def pdf(alpha, x, theta=.5):
return alpha * theta ** alpha / x ** (alpha + 1)
class ParetoI:
alpha = None
theta = None
def __init__(self, alpha, theta=.5):
self.alpha = alpha
self.theta = theta
return
def pdf(self, x):
return pdf(self.alpha, x, self.theta)
Answered By - Tim Roberts
0 comments:
Post a Comment
Note: Only a member of this blog may post a comment.