Managing and Mining Graph Data part 56 is a comprehensive survey book in graph data analytics. It contains extensive surveys on important graph topics such as graph languages, indexing, clustering, data generation, pattern mining, classification, keyword search, pattern matching, and privacy. It also studies a number of domain-specific scenarios such as stream mining, web graphs, social networks, chemical and biological data. The chapters are written by leading researchers, and provide a broad perspective of the area. This is the first comprehensive survey book in the emerging topic of graph data processing. . | 5540 MANAGING AND MINING GRAPH DATA In concrete terms we compare the following five alternatives Eoim The structural PsN-scoring approach similar to 9 cf. Subsection but based on the unordered Roim_unord reduction. Esll tree The frequency-based Pfrcq-scoring approach as in 13 14 cif. Subsection based on the RSubtree reduction. Ecomb i3 The combined approach from 131 cf. Subsection based on he Roim_unord tind Rsubtree reductions. Ecomb i4 The combined approach from 14-1 cf. Subsection based on he Rsubtree reduction. Etota The combined approach as in 1 1 cf. Subsection but with the Rtotai_w reduction like n t25t but with weights and without temporal edgee cf. Subsection . We presrn the resuL s ithr number of thr lirrt position in which a bug is eouni of he five cxpcrimcntL for ail fourteen bugs in Table . We represent a bug which is not dtscovcrsd wiih the respective approach with 25 the total numier of me hods of tht program. Note that with the frequency-based and the combmed method rankings there ti tuaH y is information available where a bug is located within a methodi and tti the context of which subgraph it appears. The fohowing comparisons leave amide this additional information. Exp . Bug 1 2 3 4 5 6 7 8 9 10 11 12 13 154 Eoim 255 3 1 3 2 4 3 1 1 6 4 4 25 4 Esubtree 3 3 1 1 1 3 3 1 255 2 3 3 3 3 Ecomb 13 1 3 1 2 2 1 2 1 3 1 2 4 8 5 Ecomb 14 3 2 1 1 1 2 2 1 18 2 2 3 3 3 Etotal 1 5 1 4 3 5 5 2 255 2 5 4 6 3 Table . Experimental results. Structural Frequency-Based and Combined Approaches. Comparing the resulta from Eaim and ErL-titrcc. the frequencytbared approach Esuti treei performs almost always as good or better than the structural one Eoimi- Ttis demonstrates that maSyzing numerical ca l frequencies is adequate to locate bugs. Bugs 11 9 and 13 iLSustrate that both anproaches alone cannot find certain bugs. Bugs 9 ennot be found by comparing call frequencies ErL-titrcc . TCis is becaure Bug 9 lsa modified condition which .
