GRAPH PATTERN MINING: A SURVEY OF ISSUES AND APPROACHES
B. Bhargavi · 2012
Most of the internet data that is available in public that are analyzed/archived is graph structured in nature. Graphs form a powerful modeling tool in many areas that include chemistry, biology, www etc. Hence there is a demand for efficiently querying such large graph data. Graph pattern matching problem is to find all the patterns from a large data graph that match the given graph pattern. The survey paper discusses tree pattern matching and graph pattern matching techniques and efficient computation of compressed transitive closure using 2-hop labeling. Join based algorithms are proposed which are two step filter(R-semijoin) and fetch(R-join) steps that are implemented using cluster-based index in relational database context. Optimization techniques like R-join order selection with R-semijoin enhancement and interleaving R-joins and R- semijoins are proposed which make graph pattern matching efficient.