90 likes | 279 Views
Graph Systems. Howles. Graph Representations. Sometimes the data naturally fits the structure of a graph Nodes (vertices) Edges (links). Advantages of Graphs. Usually well understood – a typical topic in even introductory data structure and algorithms courses Some good APIs exist
E N D
Graph Systems Howles
Graph Representations • Sometimes the data naturally fits the structure of a graph • Nodes (vertices) • Edges (links)
Advantages of Graphs • Usually well understood – a typical topic in even introductory data structure and algorithms courses • Some good APIs exist • Clear understanding of the model
Types • Directed Graph (DiGraph) • Undirected Graph • The link/edges carry data about relationships between the nodes (vertices)
Typical Measures • Indegree – possible relevance if high (links in a network structure) • Outdegree – possible relevance (shows relationships • Either may have different relevance depending on whether the graph is directed or undirected
Measures • Number of paths to a given node (vertex) • Shortest path between two nodes • Coloring algorithms • Minimal spanning tree • .. And others, depending on the application
Graph Database System • Increasing applications and interest in graph-based structures (Internet, social networking interactions) have produced specialized database support • Data is modeled as nodes and edges • Nodes may have properties • An edge may have a label or role • Credit: Jennifer Widom
Examples • A graph of people (nodes) may represent one or more role (parents, friends) • Neo4j (implemented in Java) , FlockDB(Twitter), Pregel (Google)