A clear distinction is made in JAS-mine between objects with a modelling content, which specify the structure of the simulation, and objects which perform useful but auxiliary tasks, from enumerating categorical variables to building graphical widgets, from creating filters for the collection of agents to computing aggregate statistics to be saved in the output database. JAS-mine extends the Model-Observer paradigm introduced by the Swarm experience and introduces a new layer in simulation modeling, the Collector.
This three-layer methodological protocol allows for extensive re-use of code and facilitates model building, debugging and communication. Moreover, a separation is made between code and data, with all parameters and input tables stored either in an input database or in specific MS Excel files. The regression package provides tools for simulating outcomes from standard regression models (OLS, probit/logit, multinomial probit/logit): in particular, there is no need to specify the variables that enter a regression model, as they are directly read from the data files. This greatly facilitates exploration of the parameter space, testing different econometric specifications, and scenario analysis. |
The JAS-mine platform > Concept >