Controlled simulation of longitudinal processes
This project examines longitudinal network data with temporally ordered or synchronous arrows. WikiSysopWikiSysop 07:17, 20 August 2008 (PDT)
Generations are identified from time-depth partitions. Synchronies are within generations.
Baseline models are constructed by detaching forward arrows from their future nodes within each generation (descendant or parental) and rewiring each to one of the vacated nodes. The reattachments are done 100-1000 times to build Monte Carlo distributions on some parameters.
Null models are tested as departures from the means of the MC distributions
Inferential models by specifying the parameters of theoretical distributions, generating these distributions 100-1000 times, and testing fit to the means and variances of the MC model distributions.
This can be done with
- Biotech industry data (given that the link attractors are identified from previous study)
- Industrial hierarchy data, Tsutomu Nakano
- All the kinship datasets: See Generation depth partition and Software: Kinship simulation
- Historical data on new road construction or upgrades