What are the Mining Graphs and Networks?

Graphs defines a more general class of mechanism than sets, sequences, lattices, and trees. There is a wide range of graph applications on the internet and in social networks, data networks, biological web, bioinformatics, chemical informatics, computer vision, and multimedia and content retrieval. The applications of mining graphs and networks are as follows −

Graph Pattern Mining − It is the mining of frequent subgraphs in one or a set of graphs. There are various approaches for mining graph patterns can be categorized into Apriori-based and pattern growth–based approaches.

It can mine the set of closed graphs where a graph g is closed if there continue no suitable supergraph g’ that produce the similar support count as g. Furthermore, there are several variant graph patterns, such as approximate frequent graphs, coherent graphs, and dense graphs. User-defined constraints can be driven deep into the graph pattern mining phase to enhance mining efficiency.

Statistical Modeling of Networks − A network includes a set of nodes, each equivalent to an object related to a set of properties, and a set of edges (or links) linking those nodes, describing relationships between objects.

A network is homogeneous if some nodes and links are of the similar type, including a friend network, a coauthor network, or an internet page network. A network is heterogeneous if the nodes and connection are of different types, including publication networks (connecting authors, conferences, papers, and text), and health-care networks (connecting doctors, nurses, patients, diseases, and treatments).

Data Cleaning, Integration, and Validation by Information Network Analysis − Information redundancy can exist between the several elements of data that are interconnected in a huge network. Information redundancy can be analyzed in such networks to implement quality data cleaning, data integration, data validation, and trustability search by network analysis.

Clustering and Classification of Graphs and Homogeneous Networks − Cluster analysis methods have been produced on huge networks to uncover network mechanism, find hidden communities, hubs, and outliers depends on network topological mechanism and their related properties. There are several types of network clustering methods have been produced and can be categorized as partitioning, hierarchical, or density-based algorithms.

Clustering, Ranking, and Classification of Heterogeneous Networks − A heterogeneous network includes interconnected nodes and connection of multiple types. Such interconnected mechanism include rich data, which can be used to mutually improve nodes and links, and propagate observation from one type to another.

Clustering and ranking of such heterogeneous web can be implemented closely associated in the context that highly ranked nodes in a cluster can contribute more than their lower-ranked match in the computation of the cohesiveness of a cluster.