Hands-On Artificial Intelligence for Cybersecurity

K-Means pros and cons

Among the advantages of the K-Means algorithm we can remember, in addition to its simplicity of use, its high scalability makes it preferable in the presence of large datasets.

The disadvantages instead are essentially due to the inappropriate choice of the k parameter, representative of the number of clusters, which, as we have seen, requires particular attention on behalf of the analyst, who will be called to carefully evaluate this choice on the basis of an EDA, or proceeding by trial and error.

Another disadvantage associated with the using the K-Means algorithm is determined by the fact that it provides poorly representative results in the presence of datasets characterized by high dimensions.

As a result, the phenomenon ...