### Table 1. assembly of grains graph index number of elements

"... In PAGE 21: ...2 Graph models of heterogeneous media Let us pursue the planar graph representation of granular media in some more detail, Satake @71,72#. First, we list in Table1 a correspondence between a system of rotund grains and its graph model. Besides the vertex and edge sets intro- duced earlier, we also have a loop set L.... ..."

Cited by 2

### Table 3.6 4 Edge-Forwarding Index of Frobenius Graphs of p-

### TABLE II The augmented valences of the symmetry non-equivalent vertices of the nine graphs depicted in Figure 2 and the values of the AVC(1/2d) index for the graphs

1998

### TABLE 1. The AC index of dissimilarity between five graph distance measures and the geometric distance, computed for the 76 benzenoid hydrocarbons from Figure 11

2001

"... In PAGE 21: ...271 TABLE1 . (Continued) 43 0.... ..."

### Table 2. Characteristics of various databases used for the scalability experi- ment. This table shows the number of graphs in each database, the average number of nodes and edges per graph in the databases, and the number of entries in the FragmentIndex.

### Table 3: Normalized Performance of HiLog of indexing can be solved by using XSB apos;s transformational indexing. The discrimination graph for this fragment is shown in gure 4, and is essentially a union of the graphs of the predicates in the fragment.

1994

"... In PAGE 13: ... In this form, the path(G)/2 predicate is not much less e cient than if it were written in rst- order syntax; there is only some minimal overhead for the extra argument (G) that makes this predicate generic. Table3 provides performance results for two sets of example programs. The rst set consists of the naive reverse Prolog benchmark (nrev), and a database program (query) that searches populations and areas to nd countries of approximately equal population density.... In PAGE 13: ... p/2 took the form of a chain. Table3 presents relative times for the HiLog encoded programs with and without specialization. HiLog syntax provides an elegant way to construct and manipulate sets.... ..."

Cited by 191

### Table 5 gives two interesting patterns extracting from bar graph where different shades represents different class labels. We argue Pattern B is more useful than Pattern A, because we can easily find a split point, which is located on the boundary between the fifth cell and the sixth cell. In order to help users to evaluate the created bar graph, we design a novel function, called Panda Index, to evaluate the purity of bar graph where the stronger the relationship two attributes have, the purer the created sub-bar graphs will be. Suppose L is the length of the generated bar graph. The panda index for a bar graph is defined as:

"... In PAGE 5: ... Because these numbers represent purity by counting the cell weight, the larger the number, the better. Table5 . Two Interesting Patterns Pattern A Pattern B Table 6.... ..."

### Table 5. Fuzzy graph results

"... In PAGE 10: ....124 0.36 0.206 0.34 After conceptual graph production, we compare concept index (Table 3), relation index (Table 4) and FCG index ( Table5 ). Results show that concept indexes perform better relation ones.... ..."

### Table 3. Comparison with 3 well-known indexes: the Calinski and Harabasz (CH) index; Krzanowski and Lai (KL) index; and Hartigan (Hart) index with 3 well-known clustering algorithms: bisecting k-means, graph-based, and hierarchical { a (p) indi- cates a correct estimation.

2004

Cited by 1