Basics

The most basic thing you can use pykelihood for is creating and manipulating distributions as objects. The distribution parameters can be accessed like standard Python attributes. Sampling from the distribution or computing the quantiles can be done using the same semantics as with the Python package scipy.stats.

>>> from pykelihood.distributions import Normal
>>> n = Normal(1, 2)
>>> n
Normal(loc=1.0, scale=2.0)

n is an object of type Normal. It has 2 parameters, loc and scale. They can be accessed like standard Python attributes:

>>> n.loc
1.0

Using the Normal object, you can calculate standard values using the same semantics as scipy.stats:

>>> n.pdf([0, 1, 2])
array([0.17603266, 0.19947114, 0.17603266])
>>> n.cdf([0, 1, 2])
array([0.30853754, 0.5       , 0.69146246])

Or you can also generate random values according to this distribution:

>>> n.rvs(10)
array([ 3.31370986,  5.02699468, -0.3573229 ,  1.00460378, -3.26044871,
        1.86362711, -0.84192901,  0.81132182, -2.03266978,  1.48079944])