Applications of Social Graphs: Author, Email, Telephone Using Graph Mining Techniques

Applications of Social Graphs: Author, Email, Telephone Using Graph Mining Techniques

Bapuji Rao, Sasmita Mishra, Sarojananda Mishra
Copyright: © 2018 |Pages: 35
DOI: 10.4018/978-1-5225-3646-8.ch013
OnDemand:
(Individual Chapters)
Available
$37.50
Current Special Offers
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

The chapter focuses on the applications of three social graphs, namely author, email, and telephone. The important application of author graph is the real network communications among the authors publishing research papers. The authors have proposed an algorithm to extract sub-graphs of authors. Individual email sending is more frequently within a group rather than group sending. Such group entities can be detected and analyzed further for knowledge discovery. For this, the authors have proposed an algorithm for extraction of strong and weak communicator's sub-graph. The telephone graph with multiple distinct types of connectivity information can be derived as layers and considered an important application. Therefore, the information retrieval can be represented as multi-layer graphs. Each layer has its own set of edges over the same underlying nodes of base-layer. Therefore, the edges of different layers typically are related but unique in behavior. For this, the authors have proposed an algorithm for representation of telephone graph as a multi-layer graph.
Chapter Preview
Top

Literature Review

Data Mining is the process of discovering interesting, non-trivial, implicit, previously unknown and potentially important information and patterns from data by Cook et al (2007). A graph can represent any complex relationships between data objects in pictorial way. The entities are mapped to vertices and the relationship is represented as an edge between the related pair of vertices. Hence graph based data mining is called as graph mining which aims at discovering the interesting sub-structures within a structural data. The authors have proposed an algorithm for discovering sub-graphs such as individual authors published papers, authors associated with individually published papers and authors associated with others having no individual published papers from authors graph using graph mining techniques.

Email mining includes spam detection, email categorization, contact analysis, email network property analysis and email visualization proposed by Tang et al (2014). Graph based ranking algorithms for email correspondents according to their degree of expertise on subjects of interests proposed by Dom et al (2003). A graph based mining approach for email and document classification proposed by Aery et al (2014). Hence the authors have proposed an algorithm for discovering sub-graphs such as weak connected and strong connected from email graphs using graph mining techniques.

Complete Chapter List

Search this Book:
Reset