|Saturday, February 17|
|CS23 Data Science Applications||
Sat, Feb 17, 11:00 AM - 12:30 PM
Recent Advances in the Analysis and Detection of Communities in a Network (303525)*Frederick Kin Hing Phoa, Institute of Statistical Science, Academia Sinica
Keywords: Network Analysis, Community Detection, Scan Statistics, Network Centrality
Networks are widely applied to explain relationship among the nodes or actors of interest, including but not limited to, members of a club, protein interaction in a neural system, food chain web in an ecological system, and many others. In general, network data analyses include the detection of network community, the identification of network center node, the characterization of network property and so on. However, many classical methods in the analysis of network data, especially social networks, lack a prompt statistical verification and inference. This talk introduces several recently developed statistical methods in network data analysis that fully embrace statistical inference via likelihood approaches. These methods are demonstrated to real-life networks. Numerical simulations and real-life applications show that these methods are promising and perform well in terms of accuracy and efficiency.