Abstract # 2:

Scheduled for Wednesday, June 17, 2015 09:00 AM-12:00 PM: (COCC Science Center ) Pre-Conference Workshop


B. McCowan1,2, B. Beisner1,2, J. Jin1,2, J. Vandeleest1,2, K. Fuji3, S. Chan3, H. Wang3, A. Shev3 and F. Hsieh3
1UC Davis, Population Health & Reproduction, School of Veterinary Medicine, Davis, CA 95616, USA, 2CNPRC, 3Department of Statistics, UC Davis
     The McCowan Animal Behavior Laboratory in collaboration with the Statistics Department at UC Davis is offering a 3-hour workshop on our newly developed network analytical approaches. These methods are data-driven network algorithms in R developed for looking at patterns in dominance and subordination, grooming, and how patterns in relationships map to individual and group level attributes (e.g., personality). Our research team will provide R code and/or packages for these methods and instruct members how to use them. We will offer three different methods: (1) percolation-conductance using a beta random field finds multi-step, directed pathways between all pairs of nodes to assess the global directional consistency of information flow between dyads; (2) data-cloud-geometry performs a random walk to identify clusters of nodes that are more similar to one another than they are to other clusters; and (3) joint-network-modeling examines the interdependent relationship between two or more behavioral (biological) networks containing the same nodes. Participants are expected to bring their own laptop computers with R downloaded. Our research team will provide information on how to prepare and load network datasets in R, as well as sample datasets for participants to practice with. However, participants are encouraged to bring their own network data to the workshop. This workshop likely will be of interest to ASP members studying all aspects of social behavior in study groups of >10 individuals.