add jitter to duplicated values for continuous splitting rule

pymc_bart/pgbart.py

@@ -155,10 +155,9 @@ def __init__(
         else:
             self.split_rules = [ContinuousSplitRule] * self.X.shape[1]
 
-            jittered = np.random.normal(self.X, self.X.std(axis=0) / 12)
-            min_values = np.min(self.X, axis=0)
-            max_values = np.max(self.X, axis=0)
-            self.X = np.clip(jittered, min_values, max_values)
+        for idx, rule in enumerate(self.split_rules):
+            if rule is ContinuousSplitRule:
+                self.X[:, idx] = jitter_duplicated(self.X[:, idx], np.std(self.X[:, idx]))
 
         init_mean = self.bart.Y.mean()
         self.num_observations = self.X.shape[0]
@@ -693,6 +692,21 @@ def inverse_cdf(
     return new_indices
 
 
+@njit
+def jitter_duplicated(array: npt.NDArray[np.float_], std: float) -> npt.NDArray[np.float_]:
+    """
+    Jitter duplicated values.
+    """
+    seen = []
+    for idx, num in enumerate(array):
+        if num in seen:
+            array[idx] = num + np.random.normal(0, std / 12)
+        else:
+            seen.append(num)
+
+    return array
+
+
 def logp(point, out_vars, vars, shared):  # pylint: disable=redefined-builtin
     """Compile PyTensor function of the model and the input and output variables.
 

pymc_bart/split_rules.py

@@ -79,7 +79,7 @@ def divide(available_splitting_values, split_value):
 class SubsetSplitRule(SplitRule):
     """
     Choose a random subset of the categorical values and branch on belonging to that set.
-    This is the approach taken by  Sameer K. Deshpande.
+    This is the approach taken by Sameer K. Deshpande.
     flexBART: Flexible Bayesian regression trees with categorical predictors. arXiv,
     `link <https://arxiv.org/abs/2211.04459>`__
     """