A Case of Study with the Clustering R Library to Measure the Quality of Cluster Algorithms.

This pattern has been addressed by researchers, as it allows us to categorize unlabelled data, and its use is suitable for automatic data classification to reveal concentrations of data. While many classification methods have been proposed, there are no criteria on which methods are more suitable for a given dataset. This paper presents the Clustering library which contains a set of well-known clustering algorithms to cover two objectives: first, grouping data in a homogeneous way by establishing differences between clusters; and second, generating a ranking between algorithms and the attributes analyzed in the dataset. Finally, through the GUI we can run the experiment without knowing the library parameters.