Software Framework for Data Analysis of Graph-Based Social Networks
Chapter One
Objectives of the Study
- To develop a robust software framework for graph-based social network data analysis.
- To design models that can manage space and time complexity in graph structures efficiently.
- To propose techniques for effective sampling of large-scale graph networks using Kronecker double cover and curvature measures.
- To implement entity detection using a CRF model enriched with semantic, linguistic, and domain-specific features.
- To explore the integration of attention-based deep learning models with domain knowledge for improved classification and learning.
CHAPTER TWO
LITERATURE REVIEW
Social Networks and Graphs
Social networking is a social structure composed of individuals or organizations which are called “nodes” connected by one or more types of relationships of interdependence, for example, friendship, belief or work. There are several works developed in the area of social network analysis. Among them, we highlight those related to the area of exact sciences: network theory and graph theory [2,3,16], which established several metrics to analyze different characteristics of a social network.
One of the most common methods of computational representation of social networks is the use of graphs. Graphs are data structures that have been widely studied by computer science and applied to represent problems of several domains. A graph consists of a set V of vertices (or nodes) and a set E of edges. Each edge connects two vertices. A graph may or may not be directed. In a directed graph (also called digraph), the edges (also called arrows) have a direction, which means that the edge parts from the node A and reaches the node B (and this edge would be different from one which parts from the node B and reaches the node A). In an undirected graph there is no order relation between the nodes connected by an edge [5].
One of the ways to use graphs to represent social networks is to consider that each individual within a network is a node and each link (or relation) between individuals is an edge. The decision whether or not the graph is directed depends on the type of relationships between individuals that one wants to represent. For instance, the relationship of co-authorship does not require the use of directed edges: if the researcher A is the co-author of a paper with researcher B then, necessarily, B is also co-author of a paper with researcher A. On the other hand, the advising relationship (advisor / student) is a relationship that requires a direction: if the researcher A is the advisor of the student B, it does not mean that the student B is the advisor of the researcher A.
There are several graph-related concepts that can be useful for the study and analysis of social networks. These concepts are based on definitions from the work of [6], [11] and [15].
The focus on specific parts of a graph leads to the concept of a subgraph. A subgraph is a specific group of nodes and edges from the original graph. A connected component of a graph is a subgraph where all nodes can reach each other.
There is a difference in the notion of node degree, which depends on whether the graph is directed or undirected. In an undirected graph, the degree of a node is the number of edges connected to it. In a directed graph, this concept is divided into two: indegree and outdegree. The degree of a node can vary from 0 (when the node is isolated) to g-1 meaning the node is connected to all nodes in the graph (where g is the number of nodes in the graph) disregarding possible self-loops.
CHAPTER THREE
THEORETICAL FRAMEWORK
The theoretical foundation of this research draws from a blend of computational theories, mathematical models, and machine learning paradigms. The key underpinning theories and principles guiding this study include:
Graph Theory
At the heart of social network analysis lies graph theory, which provides the language, models, and mathematical structures needed to represent users (nodes) and relationships (edges). The study leverages foundational concepts such as adjacency matrices, degree distributions, graph traversal, and centrality measures. Key theoretical constructs applied include:
REFERENCES
- P. Borgatti,M. G. Everett, L. C. Freeman (2006) UCINET 6 for Windows – software for social network analysis: User‟s guide. Analytic Technologies.
- Cormen,C. E. Leserson, R. L. Rivest (2001). Introduction to Algorithms.The MIT Press. Second Edition.
- Diestel (2006). Graph Theory. Springer-Verlag, Third edition.
- C. Freeman. Centrality in social networks: Conceptual clarification. Social Networks, 1(3), pages 215-239.
- A. Hanneman, M. Riddle. (2005). Introduction to social network methods. Riverside, California.
- P. Mena-Chalco, R. M. Cesar Junior (2009). scriptLattes: an open-source knowledge extraction system from the Lattes platform. Journal of the Brazilian Computer Society, 15(4):31-39.
- E. J. Newman (2001). The structure of scientific collaboration networks. Proceedings of The National Academy of Sciences of USA (PNAS), volume 98, pages 404-409.
- E. J. Newman (2003). The structure and function of complex networks. SIAM Review, 45(2), pages167-256.
- Otte, R. Rousseau (2002).Social network analysis: a powerful strategy, also for the information sciences. Journal of Information Science, 28(6), pages 441-453.
- A. Población, R. Mugnaini, L. M. S. V. C. Ramos (2009). Redes sociais e colaborativas em informação científica. Editora Angellara, São Paulo.
