Clustering Subjects by Quotational Prominence (Bible Part and Book)
The following 3D plot shows all subjects clustered using K-Means by the outgoing reference prominence of the parts (Old Testament, New Testament, and Apocrypha) and Bible books
of their scriptural quotations; parts and books are separate for each Bible version. Hover over the points to view names and cluster ids. Each cluster center,
labeled with the numerical weights of the cluster's most
defining features, is marked with a colored 'x'; the higher this weight for a feature, the more prominently that feature (Bible part or book)
appears for the referencing entities in its cluster (in this case, the referencing entities are the subjects).
For a clearer view, zoom and drag areas of the plot.
To search for subjects with a similar reference patterns, search for an author and toggle the number of nearest neighbors.
The subjects are each represented by a vector that is 353 features long, and its dimensionality has been reduced using Principal Component Analysis for visualization.
