Social Big Data Analytics: Practices, Techniques, and Applications

✍ Scribed by Bilal Abu-Salih, Pornpit Wongthongtham, Dengya Zhu, Kit Yan Chan, Amit Rudra

• Publisher: Springer
• Year: 2021
• Tongue: English
• Leaves: 226
• Edition: 1st ed. 2021
• Category: Library

✦ Synopsis

This book focuses on data and how modern business firms use social data, specifically Online Social Networks (OSNs) incorporated as part of the infrastructure for a number of emerging applications such as personalized recommendation systems, opinion analysis, expertise retrieval, and computational advertising. This book identifies how in such applications, social data offers a plethora of benefits to enhance the decision making process.

This book highlights that business intelligence applications are more focused on structured data; however, in order to understand and analyse the social big data, there is a need to aggregate data from various sources and to present it in a plausible format. Big Social Data (BSD) exhibit all the typical properties of big data: wide physical distribution, diversity of formats, non-standard data models, independently-managed and heterogeneous semantics but even further valuable with marketing opportunities.

The book provides a review of the current state-of-the-art approaches for big social data analytics as well as to present dissimilar methods to infer value from social data. The book further examines several areas of research that benefits from the propagation of the social data. In particular, the book presents various technical approaches that produce data analytics capable of handling big data features and effective in filtering out unsolicited data and inferring a value. These approaches comprise advanced technical solutions able to capture huge amounts of generated data, scrutinise the collected data to eliminate unwanted data, measure the quality of the inferred data, and transform the amended data for further data analysis.  Furthermore, the book presents solutions to derive knowledge and sentiments from BSD and to provide social data classification and prediction. The approaches in this book also incorporate several technologies such as semantic discovery, sentiment analysis, affective computing and machine learning.

This book has additional special feature enriched with numerous illustrations such as tables, graphs and charts incorporating advanced visualisation tools in accessible an attractive display.

✦ Table of Contents

Preface
Contents
Chapter 1: Social Big Data: An Overview and Applications
1.1 Introduction
1.2 SBD: An Overview
1.2.1 Definition of SBD
1.2.2 Types of Social Data Services
1.3 SBD Applications and Analytics
1.3.1 Credibility of SBD
1.3.2 Domain of Interest in SBD
1.3.3 SBD Predictive Analytics
1.3.4 Affective Design in the Era of SBD
1.3.5 Social Sentimental Analysis
1.4 Conclusion
References
Chapter 2: Introduction to Big Data Technology
2.1 Introduction
2.2 History of Big Data
2.3 Characteristics of Big Data
2.4 Cloud Computing
2.4.1 Introduction to Cloud Computing
2.4.2 Cloud Computing Service Models
2.4.3 Cloud Computing Deployment Models
2.4.4 Brief Introduction to Amazon AWS, Microsoft Azure, and Google Cloud Platform
2.5 Could Data Lakes - Snowflakes
2.6 Enterprise Data Center/Cloud - Cloudera Data Platform
2.6.1 Overview of Cloudera Data Platform
2.6.2 Hadoop HDFS
2.6.3 Apache Hadoop YARN (Yet Another Resource Negotiator)
2.7 Apache Spark
2.7.1 What Is Apache Spark
2.7.2 Main Spark Components
2.8 Apache HBase
2.8.1 Concepts in HBase
2.8.2 Apache HBase Architecture
2.8.3 Choosing a Row Key for a HBase Table
2.8.4 Basic Operations on HBase Table from HBase Shell
2.8.5 Comparison of HBase with RDMBS
2.9 Apache Solr
2.9.1 What Is Solr
2.9.2 Features of Solr
2.9.3 How Solr Works
2.9.4 SolrCloud Setup - A Tutorial
2.10 Resources
2.11 Conclusion
References
Chapter 3: Credibility Analysis in Social Big Data
3.1 Introduction
3.2 An Overview of Credibility in SBD
3.3 Credibility Approaches in SBD
3.3.1 Generic-Based Trustworthiness Approaches
3.3.2 Domain-Based/Topic-Specific Trustworthiness Approaches
3.3.3 Assessment of Approaches Incorporating Trust in SBD
3.4 Case Study on Social Credibility Analysis
3.4.1 Tweets Acquisition and Pre-Processing
3.4.1.1 Dataset Selection
3.4.1.2 Tweets Pre-Processing
Data Integration and Temporary Storage
Data Cleansing
Data Storage
3.4.2 Domain Classification & Sentiment Analysis
3.4.3 Features Extraction and Selection
3.4.3.1 Topic Distinguishing Mechanism
3.4.3.2 Users´ Metadata Analysis
3.4.3.3 Features Extraction
3.4.4 Experimental Results
3.4.4.1 Dataset Selection and Ground Truth
3.4.4.2 System Evaluation
3.5 Conclusion
References
Chapter 4: Semantic Data Discovery from Social Big Data
4.1 Introduction
4.2 Semantic Analysis
4.2.1 Ontology: Origin and Definition
4.2.2 Social Media Services Incorporating Semantic Analysis
4.3 Domain Knowledge Modelling, Inference and Storage
4.3.1 Annotation and Enrichment
4.3.2 Interlinking
4.3.3 Semantic Repository
4.3.4 Politics Ontology
4.4 Semantic Analysis Tools and APIs for SBD
4.4.1 IBM Watson Cognitive Services
4.4.2 Lexalytics Intelligence Platform
4.4.3 Cogito Discover
4.5 Knowledge Graphs for SBD
4.5.1 Knowledge Graph - An Overview
4.5.2 SBD Applications Using KGs
4.5.2.1 Information Retrieval Systems
4.5.2.2 Recommender Systems
4.5.2.3 Domain Specific Applications
4.6 A Case Study on Semantic Analysis of Social Politics Data
4.7 Conclusion and Future Works
References
Chapter 5: Predictive Analytics Using Social Big Data and Machine Learning
5.1 Introduction
5.2 Predictive Modelling Framework for Social Big Data
5.3 Machine Learning Algorithms for SBD Predictive Analytics
5.3.1 Logistic Regression
5.3.2 Generalized Linear Regression
5.3.3 Naïve Bayes (NB)
5.3.4 Decision Tree Based Classification
5.3.5 Random Forest
5.3.6 Gradient Boosted Tree
5.3.7 Deep Learning
5.4 Predictive Analytics Applications, Tools and APIs for SBD
5.4.1 Applications on Incorporating Predictive Analytics for SBD
5.4.1.1 User Modelling and Personalization
5.4.1.2 Spam and Social Influence Prediction
5.4.1.3 Content Segmentation and Classification
5.4.1.4 Customer Engagement
5.4.2 Predictive Analytics Tools and APIs
5.4.2.1 RapidMiner Studio
5.4.2.2 SAS Visual Data Mining and Machine Learning
5.4.2.3 TIBCO Data Science
5.4.2.4 H2O Driverless AI
5.5 Case Study on Social Politics Domain
5.5.1 Data Generation and Acquisition
5.5.2 Dataset Pre-Processing
5.5.3 Feature Engineering and Selection
5.5.3.1 Political Domain Knowledge Inference
5.5.3.2 User Features
5.5.4 System Evaluation
5.5.4.1 Ground Truth
5.5.4.2 Experimental Settings
5.5.4.3 Experimental Results
5.5.4.4 A Comparison with LDA and SLA
5.6 Conclusion and Future Works
References
Chapter 6: Affective Design Using Social Big Data
6.1 Introduction
6.2 Affective Design Using Big Data
6.3 Analysis Affective Big Data Using Machine Learning
6.3.1 Development of Affective Model Using Big Data
6.3.1.1 Statistical Method
6.3.1.2 Fuzzy Regression
6.3.1.3 Fuzzy Neural Networks and Fuzzy Expert Systems
6.3.1.4 Neural Network
6.3.1.5 Support Vector Regression (SVR)
6.3.2 Associate Rules Between Affective Customer Needs and Perceptual Design Elements
6.3.2.1 Associate Rule Mechanisms
6.3.2.2 Uncertainty Indication in Associate Rules
6.3.3 Determination of Design Attribute Settings for Affective Design
6.4 Conclusion and Further Works
References
Chapter 7: Sentiment Analysis on Big News Media Data
7.1 Introduction
7.2 Background Knowledge
7.2.1 Text Document Classification
7.2.2 Neural Networks
7.2.2.1 Single Neuron
7.2.2.2 Activation Functions
7.2.2.3 Loss Functions
7.2.2.4 Gradient Descent
7.2.3 Deep Neural Architecture - Recurrent Neural Networks
7.2.4 Long Short-Term Memory
7.3 Word Embedding
7.3.1 Word2vec
7.3.2 Global Vectors (GloVe)
7.4 Deep Learning for Sentiment Analysis
7.4.1 DNN for NLP Tasks
7.4.2 Recurrent Neural Networks (RNN) for Sentiment Analysis
7.4.3 PMI for Sentiment Analysis
7.4.4 Sentiment Analysis with Stanford CoreNLP
7.4.5 Sentiment Analysis with Variant Word Embeddings
7.5 Sentiment Analysis Example - Property Market Sentiment of Australian News Media
7.5.1 Hadoop Cluster at Big Data Laboratory
7.5.2 Nutch as a Web Crawler to Collect Data from the Internet
7.5.3 Use Apache Solr as Search Engine to Index Data Crawled from the Internet
7.5.4 Data Set Collected for Sentiment Analysis
7.5.5 Stanford CoreNLP
7.5.6 Sentiment Analysis Using Stanford CoreNLP
7.6 Experimental Results
7.6.1 Sentiment of all Articles
7.6.2 Sentiment of Major Data Sources
7.7 Conclusion
References