diff --git a/docs/source/tutorial_tep.ipynb b/docs/source/tutorial_tep.ipynb
index a64d667..6375418 100644
--- a/docs/source/tutorial_tep.ipynb
+++ b/docs/source/tutorial_tep.ipynb
@@ -204,7 +204,7 @@
    "source": [
     "## Importing data\n",
     "\n",
-    "The TEP simulation data originally provided by [Chiang et al. (2000)](https://books.google.com/books?hl=pt-BR&lr=&id=G71zWeHzg2QC&oi=fnd&pg=PA1&ots=jTMxhd5OsG&sig=Kp14U3gjLfq8DxKi0Gw5Tpb8RaQ) can be imported into the library using the `load_tennessee_eastman` function. This function requires an argument to specify the nature of data: use 0 for normal operation data and values from 1 to 20 for various process disturbances. Below, we demonstrate how to import normal operation data for training purposes and data corresponding to disturbance event `IDV(1)` for testing. In this dataset, disturbances are introduced after 8 hours of operation. By convention, normal datasets start on `2020-01-01 00:00:00` and faulty datasets start on `2020-02-01 00:00:00`."
+    "The TEP simulation data originally provided by [Chiang et al. (2000)](https://link.springer.com/book/10.1007/978-1-4471-0347-9) can be imported into the library using the `load_tennessee_eastman` function. This function requires an argument to specify the nature of data: use 0 for normal operation data and values from 1 to 20 for various process disturbances. Below, we demonstrate how to import normal operation data for training purposes and data corresponding to disturbance event `IDV(1)` for testing. In this dataset, disturbances are introduced after 8 hours of operation. By convention, normal datasets start on `2020-01-01 00:00:00` and faulty datasets start on `2020-02-01 00:00:00`."
    ]
   },
   {
@@ -318,7 +318,7 @@
     "\n",
     "When PCA is applied to fault detection, the principal components are selected using the training data. These principal components are expected to explain the variability of the test data as well. If the test data's variability is not adequately explained by the principal components identified during training, the process is considered to be faulty and alarms are trigged.\n",
     "\n",
-    "For more details on the PCA model, please refer to the following article: https://doi.org/10.3390/pr12020251."
+    "For more details on the PCA model, please refer to the following article: [https://doi.org/10.3390/pr12020251](https://doi.org/10.3390/pr12020251)."
    ]
   },
   {