Social Network Analysis
Description
As social media dominates our lives in increasing intensity, the need for developers to understand the theory and applications is ongoing as well. This book serves that purpose.
Social network analysis is the solicitation of network science on social networks, and social occurrences are denoted and premeditated by data on coinciding pairs as the entities of opinion.
The book features:
* Social network analysis from a computational perspective using python to show the significance of fundamental facets of network theory and the various metrics used to measure the social network.
* An understanding of network analysis and motivations to model phenomena as networks.
* Real-world networks established with human-related data frequently display social properties, i.e., patterns in the graph from which human behavioral patterns can be analyzed and extracted.
* Exemplifies information cascades that spread through an underlying social network to achieve widespread adoption.
* Network analysis that offers an appreciation method to health systems and services to illustrate, diagnose, and analyze networks in health systems.
* The social web has developed a significant social and interactive data source that pays exceptional attention to social science and humanities research.
* The benefits of artificial intelligence enable social media platforms to meet an increasing number of users and yield the biggest marketplace, thus helping social networking analysis distribute better customer understanding and aiding marketers to target the right customers.
Audience
The book will interest computer scientists, AI researchers, IT and software engineers, mathematicians.
More details
Other editions
Additional editions
Persons
Mohammad Gouse Galety, PhD, is an assistant professor in the Information Technology Department, Catholic University in Erbil, Erbil, Iraq.
Chiai Al-Atroshi is a lecturer in the Educational Counseling and Psychology Department, University of Duhok, Duhok, Iraq.
Bunil Kumar Balabantaray, PhD, is an assistant professor in the Department of Computer Science and Engineering, National Institute of Technology Meghalaya, India.
Sachi Nandan Mohanty, PhD, is an associate professor in the Department of Computer Science & Engineering at Vardhaman College of Engineering (Autonomous), Hyderabad, India.
Content
Preface xi
1 Overview of Social Network Analysis and Different Graph File Formats 1
Abhishek B. and Sumit Hirve
1.1 Introduction-Social Network Analysis 2
1.2 Important Tools for the Collection and Analysis of Online Network Data 3
1.3 More on the Python Libraries and Associated Packages 9
1.4 Execution of SNA in Terms of Real-Time Application: Implementation in Python 13
1.5 Clarity Toward the Indices Employed in the Social Network Analysis 14
1.5.1 Centrality 14
1.5.2 Transitivity and Reciprocity 15
1.5.3 Balance and Status 15
1.6 Conclusion 15
References 15
2 Introduction To Python for Social Network Analysis 19
Agathiya Raja, Gavaskar Kanagaraj and Mohammad Gouse Galety
2.1 Introduction 20
2.2 SNA and Graph Representation 21
2.2.1 The Common Representation of Graphs 21
2.2.2 Important Terms to Remember in Graph Representation 23
2.3 Tools To Analyze Network 24
2.3.1 MS Excel 24
2.3.2 Ucinet 26
2.4 Importance of Analysis 26
2.5 Scope of Python in SNA 26
2.5.1 Comparison of Python With Traditional Tools 27
2.6 Installation 27
2.6.1 Good Practices 28
2.7 Use Case 29
2.7.1 Facebook Case Study 30
2.8 Real-Time Product From SNA 32
2.8.1 Nevaal Maps 33
References 34
3 Handling Real-World Network Data Sets 37
Arman Abouali Galehdari, Behnaz Moradi and Mohammad Gouse Galety
3.1 Introduction 37
3.2 Aspects of the Network 38
3.3 Graph 41
3.3.1 Node, Edges, and Neighbors 41
3.3.2 Small-World Phenomenon 42
3.4 Scale-Free Network 43
3.5 Network Data Sets 46
3.6 Conclusion 49
References 49
4 Cascading Behavior in Networks 51
Vasanthakumar G. U.
4.1 Introduction 51
4.1.1 Types of Data Generated in OSNs 52
4.1.2 Unstructured Data 52
4.1.3 Tools for Structuring the Data 53
4.2 User Behavior 53
4.2.1 Profiling 54
4.2.2 Pattern of User Behavior 54
4.2.3 Geo-Tagging 55
4.3 Cascaded Behavior 56
4.3.1 Cross Network Behavior 56
4.3.2 Pattern Analysis 58
4.3.3 Models for Cascading Pattern 59
References 60
5 Social Network Structure and Data Analysis in Healthcare 63
Sailee Bhambere
5.1 Introduction 64
5.2 Prognostic Analytics-Healthcare 64
5.3 Role of Social Media for Healthcare Applications 65
5.4 Social Media in Advanced Healthcare Support 67
5.5 Social Media Analytics 67
5.5.1 Phases Involved in Social Media Analytics 68
5.5.2 Metrics of Social Media Analytics 69
5.5.3 Evolution of NIHR 70
5.6 Conventional Strategies in Data Mining Techniques 71
5.6.1 Graph Theoretic 72
5.6.2 Opinion Evaluation in Social Network 74
5.6.3 Sentimental Analysis 75
5.7 Research Gaps in the Current Scenario 75
5.8 Conclusion and Challenges 77
References 78
6 Pragmatic Analysis of Social Web Components on Semantic Web Mining 83
Sasmita Pani, Bibhuprasad Sahu, Jibitesh Mishra, Sachi Nandan Mohanty and Amrutanshu Panigrahi
6.1 Introduction 84
6.2 Background 87
6.2.1 Web 87
6.2.2 Agriculture Information Systems 88
6.2.3 Ontology in Web or Mobile Web 90
6.3 Proposed Model 90
6.3.1 Developing Domain Ontology 91
6.3.2 Building the Agriculture Ontology with OWL-DL 94
6.3.2.1 Building Class Axioms 94
6.3.3 Building Object Property Between the Classes in OWL-DL 95
6.3.3.1 Building Object Property Restriction in OWL-DL 96
6.3.4 Developing Social Ontology 97
6.3.4.1 Building Class Axioms 99
6.3.4.2 Analysis of Social Web Components on Domain Ontology Under Agriculture System 100
6.4 Building Social Ontology Under the Agriculture Domain 100
6.4.1 Building Disjoint Class 100
6.4.2 Building Object Property 103
6.5 Validation 104
6.6 Discussion 104
6.7 Conclusion and Future Work 105
References 106
7 Classification of Normal and Anomalous Activities in a Network by Cascading C4.5 Decision Tree and K-Means Clustering Algorithms 109
Gouse Baig Mohammad, S. Shitharth and P. Dileep
7.1 Introduction 110
7.1.1 Cascade Blogosphere Information 111
7.1.2 Viral Marketing Cascades 112
7.1.3 Cascade Network Building 113
7.1.4 Cascading Behavior Empirical Research 113
7.1.5 Cascades and Impact Nodes Detection 114
7.1.6 Topologies of Cascade Networks 114
7.1.7 Proposed Scheme Contributions 117
7.2 Literature Survey 118
7.2.1 Network Failures 122
7.3 Methodology 123
7.3.1 K-Means Clustering for Anomaly Detection 123
7.3.2 C4.5 Decision Trees Anomaly Detection 124
7.4 Implementation 125
7.4.1 Training Phase ZI 125
7.4.2 Testing Phase 126
7.5 Results and Discussion 127
7.5.1 Data Sets 127
7.5.2 Experiment Evaluation 127
7.6 Conclusion 127
References 128
8 Machine Learning Approach To Forecast the Word in Social Media 133
R. Vijaya Prakash
8.1 Introduction 133
8.2 Related Works 135
8.3 Methodology 135
8.3.1 TF-IDF Technique 136
8.3.2 Times Series 137
8.4 Results and Discussion 138
8.5 Conclusion 141
References 145
9 Sentiment Analysis-Based Extraction of Real-Time Social Media Information From Twitter Using Natural Language Processing 149
Madhuri Thimmapuram, Devasish Pal and Gouse Baig Mohammad
9.1 Introduction 150
9.1.1 Applications for Social Media 153
9.1.2 Social Media Data Challenges 154
9.2 Literature Survey 157
9.2.1 Techniques in Sentiment Analysis 164
9.3 Implementation and Results 166
9.3.1 Online Commerce 166
9.3.2 Feature Extraction 167
9.3.3 Hashtags 167
9.3.4 Punctuations 167
9.4 Conclusion 168
9.5 Future Scope 171
References 171
10 Cascading Behavior: Concept and Models 175
Bithika Bishesh
10.1 Introduction 175
10.2 Cascade Networks 177
10.3 Importance of Cascades 178
10.4 Purposes for Studying Cascades 179
10.5 Collective Action 179
10.6 Cascade Capacity 180
10.7 Models of Network Cascades 180
10.7.1 Decision-Based Diffusion Models 181
10.7.2 Probabilistic Model of Cascade 181
10.7.3 Linear Threshold Model 183
10.7.4 Independent Cascade Model 183
10.7.5 SIR Epidemic Model 184
10.8 Centrality 186
10.9 Cascading Failures 189
10.10 Cascading Behavior Example Using Python 189
10.11 Conclusion 192
References 202
11 Exploring Social Networking Data Sets 205
Arulkumar N., Joy Paulose, Mohammad Gouse Galety, Manimaran A., S. Saravanan and Saleem Raja A.
11.1 Introduction 206
11.1.1 Network Theory 206
11.1.2 Social Network Analysis 207
11.2 Establishing a Social Network 208
11.2.1 Designing the Symmetric Social Network 208
11.2.2 Creating an Asymmetric Social Network 210
11.2.3 Implementing and Visualizing Weighted Social Networks 212
11.2.4 Developing the Multigraph for Social Networks 213
11.3 Connectivity of Users in Social Networks 214
11.3.1 The Degree to which a Network Exists 214
11.3.2 Coefficient of Clustering 215
11.3.3 The Shortest Routes and Length Between Two Nodes 215
11.3.4 Eccentricity Distribution of a Node in a Social Network 217
11.3.5 Scale-Independent Social Networks 218
11.3.6 Transitivity 218
11.4 Centrality Measures in Social Networks 218
11.4.1 Centrality by Degree 219
11.4.2 Centrality by Eigenvectors 219
11.4.3 Centrality by Betweenness 220
11.4.4 Closeness to All Other Nodes 220
11.5 Case Study of Facebook 221
11.6 Conclusion 226
References 227
Index 229
1
Overview of Social Network Analysis and Different Graph File Formats
Abhishek B.1* and Sumit Hirve2
1Department of Mechanical Engineering, University of Applied Sciences, Emden Leer, Germany
2Department of Computer Engineering, College of Engineering Pune, Pune, India
Abstract
Evaluating the public data from person-to-person communication destinations through the social network could create invigorating outcomes and bits of knowledge on the general assessment of practically any product, administration, or conduct. One of the best and precise public notion pointers is through information mining from social networks, as numerous clients seem to state their viewpoints on the social networks. The innovation in the Internet technologies figured out how to expand action in contributing to a blog, labeling, posting, and online informal communication. Therefore, individuals are beginning to develop keen on mining these immense information assets to evaluate the viewpoints. The Social Network Analysis (SNA) is the way toward researching social designs using graph hypothesis and networks. It integrates an assortment of procedures for analyzing the design of informal organizations, in addition with the hypotheses that target describing the hidden elements and the patterns in this framework. It is an intrinsically integrative field, which initially emerged from the sectors of graph hypothesis, statistics, and sociopsychology. This chapter will cover the hypothesis of SNA, with a short prologue to graph hypothesis and data spread. Then discuss the role of Python in SNA, followed up by building and suggesting informal communities from genuine Pandas and text-based data sets.
Keywords: Data mining, SNA, viewpoint dynamics, graph hypothesis, Python
1.1 Introduction-Social Network Analysis
A network of interactions, where the nodes comprise of number of people, and the edges comprise of interaction among the people are termed as social network [1]. The numbers of social networks and the strategies to analyze them are available since the past decades [2]. Statistics, graph theory, and sociology are the basics for the development of the area of social networks and are used in number of fields, such as business, economy, and information science [3, 4]. The analysis of a social network is analogous to the analysis of a graph because of the presence of graph, like topology of the social network. Graph analysis consists of a number of strategies but is not suitable to analyze the social networks [5-7] because of its complex characteristics. A very large-sized social network comprises of millions of edges and nodes, where the node generally possess number of attributes. The complex and large graph of social network cannot be managed using the old graph analysis strategies [8].
Email network, collaboration network, and telephone network are the various types of social networks. However, recent online social networks, like Twitter, Facebook, and LinkedIn, have gained increased popularity within a short period with a greater number of users. It was found with a survey that Facebook has crossed more than 500 million users in the year 2010 [8]. Social media acts as a highly recognized platform with rich source of data assisting well in the field of marketing of various brands, responding to changes in marketing, enhancing the brands through promotion, and eventually attaining a large number of customers [9-11]. In particular, the role of social network is very important in the area of healthcare applications. As such, the healthcare sector requires discovering new traditions to control the provider practice and measure the best practices to satisfy and improve the health outcomes. Social network analysis (SNA) concentrates on evaluating the relation among individuals, who are attached by one or more knot of interdependency, like friendship, love, trust, cooperation, or communication. Social network analysis can provide imminent into evaluating and understanding the specialized networks of communication and, hence, developing effective interventions in the network to enhance the performance of the provider and eventually, the outcomes related to health [12]. The diagrammatic representation of SNA is shown in Figure 1.1.
For illustration, let us consider that the application of online social network in analyzing the contagious diseases originated with the biological pathogens, such as influenza, chickenpox, measles, and the sexually spread viruses that transfer from one person to another [13-15].
Figure 1.1 Social network analysis.
Recent studies have observed the prologue of a number of SNA models that try to clarify how opinions develop in a population [16], with the consideration of a number of social theories. These models possess a number of common characteristics with that of the spreading and epidemics. Generally, people are considered as agents with a certain state and attached by a social network. The social links is indicated using a complete graph or with more sensible complex networks. The state of the node is typically identified using the variables, which can either be discrete or continuous, with the probability to select either one or another option [17]. The nature of individuals varies with respect to time, depending on a number of update rules, mainly with the interaction of neighbors.
1.2 Important Tools for the Collection and Analysis of Online Network Data
In the recent years, the SNA has attained more concentration in various fields of research, which is because of the flexibility in operation provided by the graph theory that is involved in reducing the countless phenomena to a basic analytical form in terms of bricks and nodes. Certainly, the social relations, transportation, trading, communication strategies, and even the brain can be framed as a network and can be analyzed. This assists in the visibility of the studies related to network analysis, leading to be advantageous in education centers, academies, and universities particularly, healthcare. A number of tools were developed to make it available to a large amount of people. The SNA library and the graphical tools are made available to physicists, mathematicians, computer scientists, and so on. The SNA, being an active area of research, can also be used for unfolding human interactions and opinion diffusions. More number of dedicated tools and libraries are available even for certain peculiar applications. However, it is a time-consuming process to select the appropriate tool for a particular task, making it inconvenient for the users.
Some of the openly available tools and libraries are discussed in this section. A multilevel solution aiming on epidemic spreading simulation is represented as Network diffusion library (NDlib), which possesses a number of significant features and is available highly to the SNA practitioners as compared with other tools. Unlike other tools, the NDlib tool is accessible to technicians, like researchers, programmers, and to non-technicians, like students and analysts. NDlib helps in rectifying the drawbacks associated with the existing libraries with reduced complexity in usage. The three elements of the generic diffusion process are the graph topology, the diffusion model, and the configuration of the model.
The configuration of the model is devised in such a way to provide the final user with negligible and logical interface to choose the diffusion processes. The simulation configuration interface finally permits the user to completely indicate the three different groups of data, such as the model-specific parameters, the attributes of nodes and edges, and finally, the preliminary condition of the epidemic process. The configuration model has an important role in library logic in such a way that it concentrates on the description of the experiment, thus leading the definition of the simulation logical over all the models [18]. The next significant software package is the NetworKit [19], which generally provide the graph algorithms, and is efficient in analyzing the capabilities of the network. It involves balancing certain combination of strength with its two-layer hybrid feature aware code [12]. Figure 1.2 illustrates the SNA using Python.
Social Network Importer: The SN organization is a module for NodeXL6, which is the unrestrained Excel 2010/2007 format for dissecting organization in the well-known Excel application software circumstance. The Bernie Hogan of Oxford Internet Institute delineates the NameGen7, which is considered as the antecedent of SN organization [20].
Social Network Organization Importer: SN organization makes inquiries to Facebook Administration Programming Interface (API) and permits the extortion of inner self-organization information for a provided Facebook client. Contingent upon account protection settings for conscience and revamp, the apparatus will likewise gather Facebook portrait information and restore the 1.5 degree sense of self-organization. As per the Facebook API protocols and regulation, the information must be gathered for a conscience who has given their Facebook username and secret word, and henceforth Social Network Importer is as of now basically valuable for analysts who need to gather their own inner self-organization information or that of few members who might have to utilize NodeXL on a machine that influence scientific approaches. In contradiction, NameGen is accessible as an application of Facebook, and it has permitted the designers of NameGen to gather a...
