Lines Matching full:canonical

31 // An implementation of the Canonical Views clustering algorithm from
60 // canonical views clustering algorithm.
63 // interchangably.  Given a weighted Graph G(V,E), the canonical views
66 // vertex j, and C is the set of canonical views. Then the objective
67 // of the canonical views algorithm is
73 // alpha is the size penalty that penalizes large number of canonical
76 // beta is the similarity penalty that penalizes canonical views that
77 // are too similar to other canonical views.
79 // Thus the canonical views algorithm tries to find a canonical view
81 // to minimize the number of canonical views and the overlap between
86 // being chosen as a canonical view. Thus if w_i is the vertex weight
95 // as the canonical views/cluster centers, and membership is a map
115   // The minimum number of canonical views to compute.
118   // Penalty weight for the number of canonical views.  A higher
119   // number will result in fewer canonical views.
123   // canonical views.  A higher number will encourage less similar
124   // canonical views.