050055 VU Applied Social Network Analysis using Advanced Data Mining (2016W)

"Data is the new oil" is just one of the sayings that underline the rapidly increasing importance of data and their adequate analysis for today´s businesses, but also for the entire society and mankind. Areas such as finance and marketing in companies, but also health care, security, astronomy, and environmental issues will gain a lot in the future by applying the most appropriate techniques to analyze the ever-increasing flood of available data. Techniques from different disciplines, such as statistics, artificial intelligence, social network analysis, computer science, among others are used to find useful pattern in "big data". The first part of this course (week 1) will provide students with solid knowledge about state-of-the-art techniques from data mining. Real-world applications and software tools will offer the ability to apply the respective methods. A class on privacy and ethical issues will also discuss boundaries of the respective technology.

The second part of this course (week 2) is dedicated to the analysis of Social Networks that play an increasing role in our society. Facebook and Twitter are just two such internet sites where users can network. Many traditional business decisions will be influenced by social network analysis (SNA). Loan granting or marketing campaigns are just two examples. But also less traditional areas, such as e.g. investigation of organized crime can benefit from this relatively new approach. The second part of this course first lays the foundation for social network analysis. Then the main topics related to SNA will be introduced. Applications with real-world data from social networks using the respective software tools will conclude the course.

Course syllabus

Week 1:
KDD process, data preprocessing, feature selection, data mining methods (clustering, neural networks, decision trees, Support Vector Machines), applications (marketing, fraud detection, risk analysis), privacy and ethical issues, current research in data mining (dynamic data mining, combination with game theory).

Week 2:
Representation of social networks, network centrality metrics, community detection, applications (churn detection in telecommunication, gang crime), classification and prediction in social networks, privacy issue in social network analysis.

Course objectives

This course enables the participants to analyze data effectively in different situations; special emphasis will be given to social network analysis. It introduces recent methodological developments from Data Mining as well as from Social Network Analysis (SNA) and presents representative applications from different areas. Students will get practical experience via group work with real-world cases as well as research insights based on discussions of relevant publications.

Instructional methods

Lectures, case discussions, group presentations, and general discussions.