"... In PAGE 11: ... We now present some representative results in tabular form. Table1 contains results from one set of 343 trials in which we computed the LMT-skeleton and the 1:17682- skeleton on the same sets of uniform random points ranging from 100 to 350 points. For the 50-point intervals, each line of data (each value of n) represents the averages of approximately 50 trials.... ..."

"... In PAGE 4: ... Therefore, only CCAs with maximum depth of 4 to 7 are considered. Width of Subgraphs: Table2 shows the average width statistics of the subgraphs for the 29 applications. A value in the table indicates the percentage of dynamic subgraphs that had an operation in that cell of the matrix layout (higher utilized cells have a darker background).... ..."

"... In PAGE 4: ... Therefore, only CCAs with maximum depth of 4 to 7 are considered. Width of Subgraphs: Table2 shows the average width statistics of the subgraphs for the 29 applications. A value in the table indicates the percentage of dynamic subgraphs that had an operation in that cell of the matrix layout (higher utilized cells have a darker background).... ..."

"... In PAGE 8: ... Sampling entire subgraphs preserves the association between each core object and all the peripheral objects neces- sary for accurate calculation of attributes. Table2 lists a generic algorithm for subgraph sampling. The algorithm first assigns sub- graphs to prospective samples, and then incrementally converts prospective assignments to permanent assignments only if the subgraphs are separated from subgraphs already assigned to samples.... In PAGE 10: ...The algorithm for subgraph sampling ( Table2 ) depends on the predicate separate(Si,Sj) which indicates whether two subgraphs consist of disjoint sets of objects. We differenti- ate among three criteria for determining subgraph separation.... ..."

"... In PAGE 3: ... This tree structure follows the recursive procedure we present in Table 2 which can be used to explore the search space for a given graph. Before we proceed to details about mining maximal frequent subgraphs, we outline the enumeration scheme discussed so far in Table1 and Table 2. Our strategy is quite straightforward: we first find all frequent trees; trees are expanded to cyclic graphs by searching their search spaces; and maximal frequent subgraphs are constructed from frequent ones.... In PAGE 5: ... Interested readers might verify that themselves. Table 3 and Table 4 integrate these optimizations into the basic enumerate technique we presented in Table1 and Table 2. Algorithm MaxSubgraph-Expansion(T) begin 1.... ..."

"... In PAGE 3: ... This tree structure follows the recursive procedure we present in Table 2. Before we proceed to details about mining maximal frequent subgraphs, we outline the enumeration scheme discussed so far in Table1 and Table 2. Our strategy is quite straightforward: we first find all frequent trees; trees are expanded to cyclic graphs by searching their search spaces; and maximal frequent subgraphs are constructed from frequent ones.... In PAGE 5: ... Interested readers might verify that themselves. Table 3 and Table 4 integrate these optimizations into the basic enumerate technique we presented in Table1 and Table 2. Algorithm MaxSubgraph-Expansion(T) begin 1.... ..."

"... In PAGE 8: ... Query terms were chosen purely based on biological interest prior to performance evaluation. Because of the small size of the focused subgraphs ( Table5 ), we could reliably generate a large number of eigenvectors. The documents in each result set were filtered for the proper data type (i.... ..."