Added Pareto RV

aesara/tensor/random/basic.py

@@ -90,6 +90,24 @@ def rng_fn(cls, rng, shape, scale, size):
 gamma = GammaRV()
 
 
+class ParetoRV(RandomVariable):
+    name = "pareto"
+    ndim_supp = 0
+    ndims_params = [0, 0]
+    dtype = "floatX"
+    _print_name = ("Pareto", "\\operatorname{Pareto}")
+
+    def __call__(self, b, scale=1.0, size=None, **kwargs):
+        return super().__call__(b, scale, size=size, **kwargs)
+
+    @classmethod
+    def rng_fn(cls, rng, b, scale, size):
+        return stats.pareto.rvs(b, scale=scale, size=size, random_state=rng)
+
+
+pareto = ParetoRV()
+
+
 class ExponentialRV(RandomVariable):
     name = "exponential"
     ndim_supp = 0

tests/tensor/random/test_basic.py

@@ -31,6 +31,7 @@
     multivariate_normal,
     nbinom,
     normal,
+    pareto,
     permutation,
     poisson,
     polyagamma,
@@ -227,6 +228,13 @@ def test_exponential_samples():
     rv_numpy_tester(exponential, test_lambda, size=[2, 3])
 
 
+def test_pareto_samples():
+    test_alpha = np.array(0.5, dtype=config.floatX)
+
+    rv_numpy_tester(pareto, test_alpha, test_fn=stats.pareto.rvs)
+    rv_numpy_tester(pareto, test_alpha, size=[2, 3], test_fn=stats.pareto.rvs)
+
+
 def test_mvnormal_samples():
     def test_fn(mean=None, cov=None, size=None, rng=None):
         if mean is None: