Merge pull request #241 from IFCA/feature-doc-label-drift

Add label drift to documentation concepts

Author: Jaime Céspedes Sisniega

docs/source/concepts.md

@@ -28,7 +28,7 @@ The different types of changes that are considered as a form of drift can be cat
 
 - **Data drift**: There is a change in ${P(X)}$. Therefore, this type of drift only focuses in the distribution of the covariates ${P(X)}$, so
 ${P_{[0, t]}(X) \neq P_{[t+1, \infty)}(X)}$. [Data drift methods](#data-drift) are designed to try to detect this type drift. Unlike *concept drift* taking place, the presence of *data drift* does not guarantee that model's performance is being affected, but it is highly probable that is happening. We have renamed *dataset shift* {cite}`rabanser2019failing` to *data drift*
-in order to maintain consistency with the *concept drift* definition.
+in order to maintain consistency with the *concept drift* definition. These *data drift* methods can also be used to detect *label drift*, also known as *prior probability shift* {cite}`storkey2009training`, where the label distribution ${P(Y)}$ is the one that changes over time, in such a way that ${P_{[0, t]}(Y) \neq P_{[t+1, \infty)}(Y)}$.
 
 ## Why do I need to use a drift detector?
 

docs/source/references.bib

@@ -117,3 +117,14 @@ @article{johari2022always
   year={2022},
   publisher={INFORMS}
 }
+
+@article{storkey2009training,
+  title={When training and test sets are different: characterizing learning transfer},
+  author={Storkey, Amos J},
+  journal={Dataset shift in machine learning},
+  volume={30},
+  number={3-28},
+  pages={6},
+  year={2009},
+  publisher={Citeseer}
+}