Syllabus
Introduction to the area:
Information retrieval, web search, social network analysis
Main areas of interest in social networks and future directions
Fundamental paradigms of social network analysis:
Data and model description, text and web page pre-processing, graph data models, static and dynamic graph properties
Subgraph isomorphism, maximum common subgraph problem
Matching and distance computation in graphs
Topological descriptors for graph structures
Frequent substructure-based transformations and mining in graphs
Graph clustering and graph classification
Techniques for web data mining:
Web crawling and resource discovery
Search engine indexing and query processing, web spamming
Ranking algorithms: PageRank and HITS
Recommender systems: content-based methods, collaborative filtering, graph-based methods, clustering methods, latent factor models
Web usage mining: data collection and pre-processing, discovery and analysis of web usage patterns
Approaches to social network analysis:
Introduction and main properties, measures of centrality and prestige
Community detection: Kernighan-Lin algorithm and its analysis, clustering algorithms, overlapping communities
Node classification and label propagation, social influence analysis, detection of experts in social network
Evolution and link prediction in social networks
Applications:
Sentiment analysis
Opinion mining - Feature-based opinion mining and summarization, opinion search and opinion spam
Data, text and multimedia information mining in social networks
Advertizing on the web
Annotation
The concept of social networks is widely used to model mutual relationships between people (but also between other objects like chemical compounds). Intriguing problems from this area range from finding important structural patterns that influence interaction among the considered actors across sentiment analysis that studies people´s opinions, emotions, and attitudes to the analysis and evolution of the network structure itself.
Recently, the trends have shifted rather towards online social networks (e.g., Facebook, LinkedIn and MySpace) which allow for efficient data collection.