DDE-1/Clustering.md

@@ -73,7 +73,7 @@ An alternative way of clustering is to treat the data space in a continuous mann
 
 With these rules, we can find out "core" particles (like dynamic stereotypes) and the neighboring particles around the core particles. The rest is treated as outliers. The algorithm first sample a point randomly from the dataset. Then, we search locally the neighboring particles. Found number of points should satisfy the first criteria. If yes, then we search for the neighbors of the neighbors, growing the cluster in the mean while. In other words, it is as if we are dropped in a data world consists of islands to be discovered. If we are dropped on an island by luck, we start exploring it and figure out its extend. If we are drop into the ocean, nothing to hold on, we trigger the emergency protocol "outlier" and re-beamed again. By doing so, we travel from one island to another, until whole sample space is visited. 
 
-If you want to learn more about the model, you can follow the details from the [original study]() and the [follow-up work]().   
+If you want to learn more about the model, you can follow the details from the [original study](https://www.aaai.org/Papers/KDD/1996/KDD96-037.pdf) and the [follow-up work](https://dl.acm.org/doi/10.1145/3068335).   
 
 ### Concept of similarity in machine learning