diff --git a/src/synthcity/plugins/privacy/plugin_privbayes.py b/src/synthcity/plugins/privacy/plugin_privbayes.py
index fd4ab8bb..d4bb6da4 100644
--- a/src/synthcity/plugins/privacy/plugin_privbayes.py
+++ b/src/synthcity/plugins/privacy/plugin_privbayes.py
@@ -153,15 +153,26 @@ def _encode(self, data: pd.DataFrame) -> Any:
         encoders = {}
 
         for col in data.columns:
-            if len(data[col].unique()) < self.n_bins or data[col].dtype.name not in [
-                "object",
-                "category",
-            ]:
+            dtype = data[col].dtype.name
+            n_unique = data[col].nunique()
+            
+            # Case 1 : categorical variables (string or category)
+            if dtype in ["object", "category"]:
                 encoders[col] = {
                     "type": "categorical",
                     "model": LabelEncoder().fit(data[col]),
                 }
                 data[col] = encoders[col]["model"].transform(data[col])
+
+            # Case 2 : discrete numerical variables with few distinct values
+            elif np.issubdtype(data[col].dtype, np.integer) and n_unique <= self.n_bins:
+                encoders[col] = {
+                    "type": "categorical",
+                    "model": LabelEncoder().fit(data[col]),
+                }
+                data[col] = encoders[col]["model"].transform(data[col])
+
+            # Case 3 : continuous variables (float or many unique values)
             else:
                 col_data = pd.cut(data[col], bins=self.n_bins)
                 encoders[col] = {