Merge pull request #220 from dariodandrea/docs/fix-typo-in-04_user_gu…

…ide.md

fix typo in 04_user_guide.md

docs/source/04_user_guide/04_user_guide.md

@@ -223,7 +223,7 @@ Once the graph has been determined, the `BayesianNetwork` can be initialised and
 
 Maximum likelihood or Bayesian parameter estimation can be used for learning the CPDs.
 > When learning CPDs of the BN,
-> - The dicscretised data should be used (either high or low granularity of features and target variables can be used);
+> - The discretised data should be used (either high or low granularity of features and target variables can be used);
 > - Overfitting and underfitting of CPDs can be avoided with an appropriate train/test split of the data.
 
 ```python