by
Aravind Kalimurthy
(Under the Direction of John A. Miller)
With data exponentially increasing in almost all fields in todays world, there comes the neces- sity of handling and querying the data efficiently. Recently, many graph databases have emerged to handle big data. This is traditionally done using regular query handling processes and subgraph isomorphism. In this research, we introduce edge labels into graph simulation algorithms, so that we can quickly query and filter graphs not only using the vertex labels but also using edge labels. We have also accommodated cardinality restrictions for edge labeled graphs that improves the quality of the search results. Query processing involves taking a query graph and trying to find its pattern in a larger data graph. We have added the capability to query the graph database using wildcards, regular expressions, and variables. This is done by replacing, in a query graph, one or more strings in edge labels with wildcards, regular expressions or variables. Experiments are done on very large graphs with up to 30 million edges.
Index words: Graph Database; Graph Simulation Algorithms; Edge Label; Cardinality; Query Processing; Widcard; Regular Expression.