Σεμινάριο Τμήματος & CEID social hour
Ημερομηνία-χώρος: Παρασκευή 1 Ιουνίου, 3-5μμ, Κτίριο Β (Αίθουσα ΑΣ)
Ομιλητής: Ανδρέας Καναβός, Μεταδιδάκτορας Ερευνητής
Τίτλος: Social Network Analysis: Defining and Evaluating Twitter Influence Metrics
Περίληψη: In this talk we will present some very popular aspects of social network analysis. The information extraction from social networks, particularly from Twitter, without precluding the application of the proposed techniques and methods to other social networks such as Facebook, are examined. Regarding social networks, individual people or groups of people are considered as nodes while the edges represent some kind of interaction among them. Specific methods and techniques for detecting communities in graphs/networks are implemented which aim at finding groups of users with certain characteristics in common. In addition, the denotation of User Influence is introduced, e.g. the nodes identification of a social network that utmost affect other nodes. Finally, the concept of Tree-Shaped Tweet Cascades, which in fact are created by the ReTweets and contribute to the provision of information diffusion of a given tweet, is presented. The talk will summarize the main concepts and for proving our findings, several results in this broad field by using real datasets will be analyzed.
This talk is based on a set of papers that were jointly conducted by A. Kanavos and several colleagues (Faculty of our Department, PhDs and MSc students). These works were published at IEEE Big Data, ICTAI, AIAI, WEBIST conferences as well as in Algorithms and Social Network Analysis and Mining (SNAM) journals.