Big Data Recommender Systems: Application Paradigms (Computing and Networks)

Cover
Contents
Foreword
1 Introduction to big data recommender systems — volume 1
	1.1 Background
	1.2 About the book
	Acknowledgments
	References
2 Theoretical foundations for recommender systems
	2.1 Introduction
		2.1.1 Definitions of RSs
		2.1.2 The need for an RS
	2.2 Applications of RSs
		2.2.1 Current use of RSs
			2.2.1.1 Product recommendation (e-commerce)
			2.2.1.2 Movie or music recommendation (entertainment)
			2.2.1.3 Scholarly search and news articles
			2.2.1.4 Services
		2.2.2 More areas for RSs
			2.2.2.1 Recommend courses for students
			2.2.2.2 Perform career counselling
	2.3 Algorithms and theoretical foundations of RSs
		2.3.1 Phases of RSs
			2.3.1.1 Information collection
			2.3.1.2 Learning phase
			2.3.1.3 Prediction/recommendation phase
		2.3.2 Types of RSs
			2.3.2.1 Content-based recommenders
			2.3.2.2 CF recommenders
			2.3.2.3 Hybrid recommenders
			2.3.2.4 Image-based recommenders
			2.3.2.5 RSs using GDBs
		2.3.3 Datasets for recommendations
			2.3.3.1 MovieLens
			2.3.3.2 Jester
			2.3.3.3 BookCrossing
			2.3.3.4 Amazon product data
			2.3.3.5 YahooWebscope datasets
	2.4 Problems related to RSs
		2.4.1 Data sparsity problem
		2.4.2 Cold start problem
		2.4.3 Scalability
		2.4.4 Overspecialization or diversity problem
		2.4.5 Vulnerable to attacks
	References
3 Benchmarking big data recommendation algorithms using Hadoop orApache Spark
	3.1 Introduction
	3.2 Big data
		3.2.1 Hadoop
		3.2.2 Presenting the MapReduce model
		3.2.3 Hadoop input/output
		3.2.4 Apache Ambari and Ambari architecture
		3.2.5 Future of Hadoop
		3.2.6 How Hadoop works in social networking
	3.3 Apache Spark
	3.4 Recommender systems
		3.4.1 Design of recommender systems
		3.4.2 Collaborative recommendation and collaborative filtering
		3.4.3 Reducing the sparsity problem
		3.4.4 Content-based recommendation
		3.4.5 Visualization of recommendation
		3.4.6 Hybrid recommendation approaches
	3.5 Systems based on nature-inspired algorithms
	3.6 Benchmarking: big data benchmarking
	3.7 Summary
	References
4 Efficient and socio-aware recommendation approaches for bigdata networked systems
	4.1 Introduction
	4.2 Background on recommendation systems and social network analysis
		4.2.1 Recommendation systems
		4.2.2 Social network analysis
	4.3 Socio-aware recommendation systems
		4.3.1 Hyperbolic path-based recommendation system
		4.3.2 Probabilistic graphical models
			4.3.2.1 Pairwise recommendation MRFs
			4.3.2.2 Preference networks
			4.3.2.3 Ordinal random fields
			4.3.2.4 Preference Markov random fields
		4.3.3 Information diffusion-aware recommendation approaches
		4.3.4 Context-based recommendations in pervasive systems
			4.3.4.1 Context-aware recommendation systems
			4.3.4.2 Context-aware recommendation systems in IoT ecosystem
			4.3.4.3 Context-aware recommendation systems and social networks
	4.4 Qualitative comparison
	4.5 Open problems and conclusion
	References
5 Novel hybrid approaches for big data recommendations
	5.1 Introduction
	5.2 Context
	5.3 The big data architecture
	5.4 Different approaches to handle big data
		5.4.1 Approaches to detect and reduce data inconsistency
	5.5 Complexity and issues of big DI
		5.5.1 Big data analysis and integration architecture
			5.5.1.1 Combine two figures at work
			5.5.1.2 Architectural patterns as development standards
	5.6 Big DI using HAs based on Fuzzy-Ontology
	5.7 Developing approaches for the crisp ontology
	5.8 Developing HAs for Fuzzy-Ontology
	5.9 Extracting the big data key business functions for the proposed HAs based on Fuzzy-Ontology
	5.10 Identify the specification for the purpose HIDAs for big data
	5.11 Real-world project: hypertension-specific diagnosis based on HIDAs
		5.11.1 Data collection
		5.11.2 Step 1: HIDAs contrivance and excellence
		5.11.3 Step 2: determine and ascertain the necessity for fuzziness in hypertension diagnosis
		5.11.4 Step 3: specify fuzzy-associated elements in hypertension data
		5.11.5 Step 4: reusing the subsisting HIDAs resources
		5.11.6 Step 5: reusing the subsisting Fuzzy-Ontology resources elements
		5.11.7 Step 6: appropriate the subsisting of Fuzzy-Ontology elements
		5.11.8 Step 7: identify appropriate Fuzzy-Ontology elements
		5.11.9 Step 8: identify appropriate crisp ontology elements
		5.11.10 Step 9: formalisation
		5.11.11 Step 10: Fuzzy-Ontology result affirmation
		5.11.12 Step 11: documentation and notes
	5.12 Mathematical simulation of hypertension diagnosis based on Markov chain probability model
	5.13 Analysis of result
	5.14 Conclusion
	References
6 Deep generative models for recommender systems
	6.1 Introduction
	6.2 Generative models
		6.2.1 Probabilistic matrix factorization
		6.2.2 Probabilistic latent semantic analysis
		6.2.3 Latent Dirichlet allocation
		6.2.4 Collaborative topic models
	6.3 Deep learning for recommender systems
		6.3.1 Restricted Boltzmann-machine-based collaborative filtering
		6.3.2 Autoencoder for recommender systems
		6.3.3 Multilayer perceptron based recommender systems
		6.3.4 RNN/LSTM for recommendation
	6.4 Deep generative models
		6.4.1 Collaborative denoising autoencoders
		6.4.2 Collaborative variational autoencoder
	6.5 Summary
	References
7 Recommendation algorithms for unstructured big data such as text, audio, image and video
	7.1 Recommender methods
		7.1.1 Content-based recommendations
		7.1.2 Collaborative recommendations
		7.1.3 Knowledge-based recommendations
		7.1.4 Demographic recommendations
		7.1.5 Hybrid recommendations
	7.2 Big data analytic
		7.2.1 Text analytics
			7.2.1.1 Steps for text analytics system
			7.2.1.2 Text recommendation using an angle-based interest model
		7.2.2 Audio analytics
			7.2.2.1 Prediction of genre-based link in a two-way graph for music recommendation
			7.2.2.2 Personalized tag-based social media music recommendation
			7.2.2.3 Graph-based quality model for music recommendation
			7.2.2.4 Music recommendation using acoustic features and user access patterns
			7.2.2.5 Learning content similarity for music recommendation
		7.2.3 Video analytics
			7.2.3.1 Real-time video-recommendation system
			7.2.3.2 Recommendation system for micro-video on big data
		7.2.4 Image analytics
			7.2.4.1 An images-textual hybrid recommendation system
			7.2.4.2 Recommendation system for styles and substitutes based on image
		7.2.5 Other recommender system
			7.2.5.1 Personalized trip advisor service
			7.2.5.2 Recommendation system with Hadoop Framework on big data
	7.3 Recommender systems: challenges and limitations
	7.4 Summary
	References
8 Deep segregation of plastic (DSP): segregation of plastic and nonplastic using deep learning
	8.1 Introduction
	8.2 Related work
	8.3 Deep learning
	8.4 Scalable architecture
	8.5 Software framework
	8.6 Software and packages
		8.6.1 TensorFlow
		8.6.2 Keras
		8.6.3 OpenCV
	8.7 Hardware components used
		8.7.1 Arduino UNO
		8.7.2 Windshield wiper motor
		8.7.3 Stepper motor
		8.7.4 Switching power supply
		8.7.5 ULN 2003
		8.7.6 Webcam
	8.8 Hardware setup for segregation
	8.9 Experiments and observation
		8.9.1 Training process
	8.10 Conclusion and future work
	AppendixA
	Appendix B
	Acknowledgments
	References
9 Spatiotemporal recommendation with big geo-social networking data
	9.1 Introduction
	9.2 Preliminaries about SAGE
	9.3 Spatial – temporal SAGE model
		9.3.1 Problem definitions
		9.3.2 Model description
		9.3.3 Model inference
		9.3.4 Spatial smoothing
		9.3.5 Parallel implementation
	9.4 Spatial item recommendation using ST-SAGE
	9.5 Experiments
		9.5.1 Experimental settings
			9.5.1.1 Datasets
			9.5.1.2 Comparative approaches
			9.5.1.3 Evaluation methods
		9.5.2 Recommendation effectiveness
			9.5.2.1 Results and analysis
			9.5.2.2 Impact of different factors
		9.5.3 Recommendation efficiency
			9.5.3.1 Model training efficiency
			9.5.3.2 Online recommendation efficiency
	9.6 Related work
	9.7 Conclusion
	References
10 Recommender system for predicting malicious Android applications
	10.1 Background
		10.1.1 Android operating system architecture
			10.1.1.1 Applications
			10.1.1.2 Application framework
			10.1.1.3 Android runtime
			10.1.1.4 Libraries
			10.1.1.5 Kernel
		10.1.2 Android application structure
		10.1.3 Application threats
	10.2 The proposed recommender system for mobile application risk reduction
		10.2.1 Preprocessing
		10.2.2 Emulation and testing
		10.2.3 Features extraction
		10.2.4 Machine learning
		10.2.5 Dataset
	10.3 Conclusion
	References
11 Security threats and their mitigation in big data recommender systems
	11.1 Introduction
	11.2 Security issues and approaches in HDFS architecture
		11.2.1 Security issues in HDFS
		11.2.2 HDFS security methods
			11.2.2.1 Kerberos construction
			11.2.2.2 Bull Eye algorithm
			11.2.2.3 Name node algorithm
	11.3 Big data recommender system attacks
		11.3.1 Attack tactics
		11.3.2 Probe attack strategy
		11.3.3 Ratings strategy
		11.3.4 Dimensions of attacks
		11.3.5 Models of attacks
			11.3.5.1 Profile injection attacks
			11.3.5.2 Push attacks
			11.3.5.3 Nuke attacks
	11.4 Recommender algorithms
		11.4.1 Association rule mining
		11.4.2 Base algorithms
		11.4.3 k-Nearest neighbor
		11.4.4 k-Means clustering
		11.4.5 Probabilistic latent semantic analysis
		11.4.6 Recommender algorithms and evaluation metrics
			11.4.6.1 User-based collaborative filtering
			11.4.6.2 Item-based collaborative filtering
			11.4.6.3 Enhanced collaborative filtering
		11.4.7 Profile classification
	11.5 Attack response and system robustness
		11.5.1 Classification of attributes
			11.5.1.1 Generic attributes
			11.5.1.2 Model-derived attributes
		11.5.2 Enhanced hybrid collaborative recommender systems
			11.5.2.1 Hybrid recommendation algorithm
			11.5.2.2 Push attacks against enhanced hybrid algorithm
		11.5.3 Defense against profile injection attacks
			11.5.3.1 Detection methods
			11.5.3.2 Detection attributes for profile classification
	11.6 Conclusion
	References
12 User\'s privacy in recommendation systems applying online social network data: a survey and taxonomy
	12.1 Introduction
	12.2 Recommender systems and techniques: privacy of online social network data
		12.2.1 Privacy: definition
		12.2.2 Online social networks, classification, and privacy
	12.3 Taxonomy of privacy
		12.3.1 Privacy concerns in social networks
		12.3.2 User-specific privacy risks and invasion
		12.3.3 Measuring privacy in online social networks
			12.3.3.1 Dichotomous approach
			12.3.3.2 Polytomous approach
		12.3.4 Privacy-preserving approaches
		12.3.5 Privacy-preserving models
			12.3.5.1 k-Anonymity
			12.3.5.2 l-Diversity
			12.3.5.3 T-Closeness
			12.3.5.4 Differential privacy
	12.4 Privacy preservation in recommender systems
	12.5 Conclusion and future directions
	References
13 Private entity resolution for big data on Apache Spark using multiple phonetic codes
	13.1 Introduction
	13.2 Related work
	13.3 Problem formulation and background
		13.3.1 Problem formulation and notation used
		13.3.2 Phonetic algorithms for privacy preserving matching
		13.3.3 The Soundex algorithm
		13.3.4 The NYSIIS algorithm
		13.3.5 Apache Spark
	13.4 A parallel privacy preserving phonetics matching protocol
		13.4.1 Multiple algorithms for phonetic matching
		13.4.2 Protocol operation
		13.4.3 Privacy discussion
	13.5 Empirical evaluation
		13.5.1 Experimental setup
		13.5.2 Experimental results
			13.5.2.1 Algorithm combination selection
			13.5.2.2 Matching accuracy
			13.5.2.3 Time performance
	13.6 Conclusions and future work
	References
14 Deep learning architecture for big data analytics in detecting intrusions and malicious URL
	14.1 Introduction
	14.2 Related works
		14.2.1 Network intrusion detection systems (NIDSs)
		14.2.2 Related works on phishing URL detection
	14.3 Background
		14.3.1 Deep neural network
		14.3.2 Recurrent neural network
		14.3.3 Convolutional neural network
	14.4 Intrusion detection
		14.4.1 Description of KDD-Cup-99 data set
		14.4.2 Description of Kyoto network intrusion detection (ID) data set
		14.4.3 Experiments on KDD-Cup-99
		14.4.4 Proposed architecture for KDD-Cup-99 data set
		14.4.5 Experiments on Kyoto network intrusion detection (ID) data set
		14.4.6 Proposed architecture for Kyoto
		14.4.7 Evaluation results for KDD-Cup-99
		14.4.8 Evaluation results for Kyoto
	14.5 Intrusion detection (ID) using multidimensional zoom (M-ZOOM) framework
		14.5.1 Density measures
		14.5.2 Problem formulation
		14.5.3 Data set description
		14.5.4 Experiments and observations
	14.6 Phishing URL detection
		14.6.1 Data set description of phishing URL detection
		14.6.2 URL representation
		14.6.3 Experiments
		14.6.4 Hyper-parameter tuning
		14.6.5 Proposed architecture for URL analysis
	14.7 Proposed architecture for machine learning based cybersecurity
	14.8 Conclusion and future work
	Acknowledgments
	References
Index
Back Cover