The analysis of the corpus of ontologies uses the values of the metrics obtained by each input ontology to provide extra information about the metrics and the corpus of ontologies.

In particular, the distribution of each metric is calculated, which gives a global overview about the behaviour of each metric in the corpus of ontologies.

Moreover, the correlation between the metrics is also calculated by using the Pearson coefficient. The results are given in a heatmap in which not significant correlations are marked with a cross.

Finally, a clustering analysis is performed by using the Evaluome framework. This framework analyses each metric separately to find the optimal number of groups of ontologies in which the ontology corpus can be splitted. For this, evaluome performs a k-means based clustering of ontologies by using each metric separately as a unique feature. Evaluome performs clusterings for a range of values of f, and returns the value of two statistical properties of the clusterings, namely, stability and goodness. The stability measures the effect of small variations on the data, and its values are in the range [0, 1], being the 1 the best result. The goodness measures how closely related the instances in a category are, and how well-separated a category is from the rest of categories, providing values in the range [-1, 1], being the 1 the best value. Given a value of k, the method requires at least k different values for the metric used for the clustering to be able to provide a result. This method suggests the optimal number of ontology clusters (k) for each metric by analyzing the values of stability and goodness.