## Demand - satisfying job for everyone ## Decision - External intervention in the parameter space (policy parameter e.g. task richness) level is possible (Yes / No). ## Draws - hierarchical factor analysis model with 200 subgroups, 40 local + 60 global coefficient (`est_param`) - U3_specify_projection: - the number of bookstore etc - U4_specify_regularize (by adding to `targe`): - difference of latent variable and observed data is added to `target` according to `family` distirbution function (negative binomial) - parameter is added to `target` accroding to prior distribution (e.g. `lognormal_lpdf()` ) - middle size (around 5 stocks) generator - bimodal parameter geometry and its mapping with tipping point boundary in phase space Examples each by Tom, Hazhir, Jeorun, David is documented in [[tipping points and bimodality]]