June 20, 2018
12:00 - 13:00
Nyquist Room - Maxell building a.164
Recent Innovations in Clustering - Discovery, Explanation and Twitter Experiments
by Ian Davidson, UC Davis - California
The idea of taking a collection of instances (or a graph) and breaking it into sub-graphs, clusters, or communities is popular in many fields with many variations. We cover novel formulations of clustering using spectral methods, quadratic programming, SAT, ILP and constraint programming, and discuss their strengths and weaknesses. We demonstrate some of our recent work on analyzing graphs of Twitter data from the 2016 US Presidential Primary Election.
Ian Davidson is full professor of artificial intelligence, machine learning and data mining at UC Davis (California). He authored more than 70 papers at the most prominent conferences in machine learning, data mining and artificial intelligence. He is well-known for his contributions to constraint- based clustering. Most recently, his algorithms have been applied in neuroscience, intelligent tutoring systems and social network analysis. Among others, he is interested in the application of constraint programming, SAT solving and Integer Programming in data mining.