Dumitrescu D., "Mathematical principles of Classification Theory ", Ed. Academiei Romane, Bucuresti, 1999  Home/Search   Document Not in Database   Summary   Related Articles

....in x, but they miss from y, v = y i (1 x i ) represents the number of index terms that exist in y, but they miss from x. It is easy to show that: s t = x and s v = y Taking account of the prior features, the meaning of the following similarity measures is easy to understand [2]: s (u v) s 2(u v) st uv st uv . 3. The criterion function Let X = be the entities set that must be classified. Our aim is to find the cluster structure of the given set. The cluster structure of the set X can be done by a partition P = A 1 , A 2 , A n ....

....in R . This point is the same with the centre of the class, as shown in Figure 1. A dissimilarity measure on X is a function D : X R, that satisfies the following axioms: a) D(x, y) # 0, b) D(x, x) 0, #x # c) D(x, y) D(y, x) X. The criterion function (J) may be defined as [2]: D(x, L i ) where D is a dissimilarity measure (for instance, a distance on R ) 4. The n mean algorithm The n mean algorithm is a very popular clustering technique. The following dissimilarity measure is considered: D(x, y) y# . The dissimilarity between a point x and the L i ....

Dumitrescu D., "Mathematical principles of Classification Theory ", Ed. Academiei Romane, Bucuresti, 1999

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC