Statistics Colloquium: STA 290
Thursday, February 20th at 4:10pm, MSB 1147 (Colloquium Room)
Refreshments at 3:30pm in MSB 4110 (4th floor lounge)
Speaker: Fushing Hsieh
University of California, Davis
Title: "Before statistics on networks"
Abstract: An undirected or bipartite binary network is viewed as a thermodynamic system. So its macrostate, minimizes the Ising energy level, is taken as information content. A new computing paradigm, Data Mechanics, is used to compute macrostate and reveals its multiscale patterns information. Network bootstrapping principle and algorithm are discussed together with the entropy evolution of bootstrapping ensembles. A directed binary network is shown as being much more complex than the previous two types. Throughout this talk real networks are illustrated. At the end I discuss which directions network modeling could possibly be realistic and practical.
NOTE: Students enrolled in BST 290 would be expected to attend this seminar