Manning, Christopher, Prabhakar Raghavan and Hinrich Schütze. 2008. “Flat Clustering” and “Hierarchical Clustering.” Chapters 16 and 17 from Introduction to Information Retrieval.
Is there a way to include more information in the clustering beyond just bag of words? for instance, can we determine whether certain topics are being discussed positively or negatively in addition to being discussed at all? Maybe within a hierarchy we might want all articles discussing topic X, and then within that cluster by positive or negative discussions of that topic. Or would that be a separate issue from the initial topic clustering?
Is there a way to include more information in the clustering beyond just bag of words? for instance, can we determine whether certain topics are being discussed positively or negatively in addition to being discussed at all? Maybe within a hierarchy we might want all articles discussing topic X, and then within that cluster by positive or negative discussions of that topic. Or would that be a separate issue from the initial topic clustering?