Hands-On Artificial Intelligence for Cybersecurity

Understanding clustering algorithms

The intuition underlying clustering algorithms consists of identifying and exploiting the similarities that characterize certain types of phenomena.

In technical terms, it is a matter of distinguishing and recognizing, within a dataset, the features whose values ​​change with high frequency, from those features whose values are shown to remain systematically stable ​​instead. Only these latter features are taken into consideration for the detection of phenomena characterized by similarity.

We can follow these two types of approaches in identifying similarities:

• Supervised: The similarities are identified on the basis of previously categorized samples (for example, the k-Nearest Neighbors (k-NNs) algorithms). ...