fix predictions when using 1 chain

Author: aloctavodia

pymc/bart/bart.py

@@ -62,11 +62,10 @@ class BART(NoDistribution):
     m : int
         Number of trees
     alpha : float
-        Control the prior probability over the depth of the trees. Even when it can takes values in
-        the interval (0, 1), it is recommended to be in the interval (0, 0.5].
+        Control the prior probability over the depth of the trees. Defaults to 0.25.
+        It is recommended to be in the interval (0, 0.5].
     k : float
-        Scale parameter for the values of the leaf nodes. Defaults to 2. Recomended to be between 1
-        and 3.
+        Scale parameter for the values of the leaf nodes. Defaults to 1.
     split_prior : array-like
         Each element of split_prior should be in the [0, 1] interval and the elements should sum to
         1. Otherwise they will be normalized.

pymc/bart/pgbart.py

@@ -14,7 +14,7 @@
 
 import logging
 
-from copy import copy
+from copy import copy, deepcopy
 
 import aesara
 import numpy as np
@@ -39,12 +39,12 @@ class PGBART(ArrayStepShared):
     vars: list
         List of value variables for sampler
     num_particles : int
-        Number of particles for the conditional SMC sampler. Defaults to 40
+        Number of particles. Defaults to 40
     max_stages : int
-        Maximum number of iterations of the conditional SMC sampler. Defaults to 100.
+        Maximum number of iterations. Defaults to 100.
     batch : int or tuple
         Number of trees fitted per step. Defaults to  "auto", which is the 10% of the `m` trees
-        during tuning and after tuning. If a tuple is passed the first element is the batch size
+        during and after tuning. If a tuple is passed the first element is the batch size
         during tuning and the second the batch size after tuning.
     model: PyMC Model
         Optional model for sampling step. Defaults to None (taken from context).
@@ -81,10 +81,10 @@ def __init__(self, vars=None, num_particles=40, max_stages=100, batch="auto", mo
         # if data is binary
         Y_unique = np.unique(self.Y)
         if Y_unique.size == 2 and np.all(Y_unique == [0, 1]):
-            self.mu_std = 6 / (self.k * self.m**0.5)
+            self.mu_std = 3 / (self.k * self.m**0.5)
         # maybe we need to check for count data
         else:
-            self.mu_std = (2 * self.Y.std()) / (self.k * self.m**0.5)
+            self.mu_std = self.Y.std() / (self.k * self.m**0.5)
 
         self.num_observations = self.X.shape[0]
         self.num_variates = self.X.shape[1]
@@ -229,8 +229,7 @@ def init_particles(self, tree_id: int) -> np.ndarray:
         Initialize particles
         """
         p = self.all_particles[tree_id]
-        particles = [p]
-        particles.append(copy(p))
+        particles = [p, p.copy()]
 
         for _ in self.indices:
             particles.append(ParticleTree(self.a_tree))
@@ -275,6 +274,9 @@ def __init__(self, tree):
         self.old_likelihood_logp = 0
         self.used_variates = []
 
+    def copy(self):
+        return deepcopy(self)
+
     def sample_tree(
         self,
         ssv,