### Table 2 The Mapping between leaf nodes and components Leaf node Component

in Safety

2007

"... In PAGE 7: ... 3. Table2 describes the mapping between the compo- nents in Fig.... ..."

### Table 2. IPv4 leaf nodes percentage

"... In PAGE 4: ...an. 02. The results are almost the same: in all these BGP tables, the leaf nodes are beyond proportion of 90% of total prefixes. Table2 shows the results. Table 2.... ..."

### Table 3: The suffixes and lcps in a leaf node.

"... In PAGE 5: ... Otherwise, it is one. Table3 shows the suffixes and lcps in a leaf node. Assume that we calculate the frequency of... In PAGE 6: ... Assume that we find patterns that occurs four times. We consider sliding window which contains the suffixes 29, 4, and 1023 in Table3 . Since s is 4, e is 5, and m is 6, 4-occurring pattern exists in this sliding window, and it is AACAGG.... ..."

### Table 6 Number of nodes to re-label in leaf node updates

"... In PAGE 21: ... Hamlet has 5 act elements. We test the following 5 cases (see Table6 and Figure 13): inserting an act element before act[1], inserting an act element before act[2], .... In PAGE 21: ...ct[2], . . . , and inserting an act element before act[5]. Table6 shows the number of nodes to re-label when applying di erent labeling schemes. In the 5 cases, V- Binary-Containment and F-Binary-Containment need to re-label many nodes (note that Hamlet has totally 6636 nodes).... In PAGE 21: ... It should be noted that V-Binary-Containment and F- Binary-Containment have one more node to re-label than BinaryString-Pre x and DeweyID(UTF8)-Pre x because act elements are the children of the root and the contain- ment schemes need to re-label the root also (modify the \end quot; value of the root). For Prime, the number of SC values that are required to re-calculate is counted in Table6 . Because Prime uses each SC value for every ve nodes [38], the number of SC values required to re-calculate is 1/5 of the number of nodes required by DeweyID(UTF8)-Pre x to re-label.... ..."

### Table 3: Number of ngrams, leaf nodes and the depth of the tree

"... In PAGE 7: ...Table 3: Number of ngrams, leaf nodes and the depth of the tree Table3... ..."

### Table 1. Pairs of images with the same leaf node keys

"... In PAGE 4: ... Therefore, the more pairs there are in the same nodes, the better would be the performance for this particular test. Results are shown in Table1 . The longer the keys are, the smaller are the sizes of image clusters in the leaf node, hence the chance of the same pairs falling into the same node is decreased.... ..."

### Table 2: Leaf node reorder algorithm test results (msecs)

2002

"... In PAGE 25: ... The query sets can be found in Appendix A. Optimization overhead given in the results speci es the time that the optimization process takes and performance gain is formulated as follows: performance gain = (processing time without opt: ; processing time with opt:) processing time without opt: : (1) The rst set of results are given in Table2 . These results show that leaf node reordering algorithm enhances the performance of the query processor.... ..."

Cited by 4

### Table 2. Numbers of distributions (leaf-nodes) and free pa- rameters.

"... In PAGE 3: ... Table 1 shows the examples of number of applied ques- tions. Table2 shows the total numbers of distributions (leaf nodes) and free parameters. From Table 1 and 2, it can be seen that FD-PDT and DS-PDT have stronger context dependencies and a higher representation ability than PDT, without increasing the number of free parameters.... ..."

### Table 3: The su xes and lcps in a leaf node.

"... In PAGE 6: ... Otherwise, it is one. Table3 shows the su xes and lcps in a leaf node. Assume that we calculate the frequency of hexa-mer.... In PAGE 6: ... Assume that we nd patterns that occurs four times. We consider sliding window which contains the su xes 29, 4, and 1023 in Table3 . Since s is 4, e is 5, and m is 6, 4-occurring pattern exists in this sliding window, and it is AACAGG.... ..."