8e46c0dfb78769a3f97db273e3c608059b74bdfa,pgmpy/factors/JointProbabilityDistribution.py,JointProbabilityDistribution,check_independence,#JointProbabilityDistribution#Any#Any#Any#,135

Before Change

        JPD = self.copy()
        if event3:
            if isinstance(event3, six.string_types):
                raise TypeError("Event 3 A list of tuples of the form (variable_name, variable_state")
            JPD.conditional_distribution(event3)
        if isinstance(event1, six.string_types):
            raise TypeError("Event 1 should be a list or array-like structure")
        if isinstance(event2, six.string_types):

After Change

                // Using the alternate definition for conditional independence
                // X and Y are conditional independent if phi(X, Z) * phi(Y, Z) is propotional
                // to phi(X, Y, Z)
                for variable_pair in itertools.product(event1, event2):
                    JPD_ev1_ev3 = JPD.marginal_distribution((variable_pair[0], event3), inplace=False)
                    JPD_ev2_ev3 = JPD.marginal_distribution((variable_pair[1], event3), inplace=False)
                    JPD_prod = JPD_ev1_ev3 * JPD_ev2_ev3
                    phi1 = Factor(JPD_prod.variables, JPD_prod.cardinality, JPD.values)
                    phi2 = Factor(JPD.variables, JPD.cardinality, JPD.values)
                    phi = phi1 / phi2
                    if(np.unique(phi.values).size != 1):
                        return False
                return True
            else:
                    JPD.conditional_distribution(event3)
        for variable_pair in itertools.product(event1, event2):

Instances