Action network: a probabilistic graphical model for social simulation
Agent-based social simulations are typically described in imperative form. While this facilitates implementation as computer programs, it makes implicit the different assumptions made, both about the functional form and the causal ordering involved. As a solution to the problem, a probabilistic grap...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Published: |
2022
|
| Online Access: | http://scholars.utp.edu.my/id/eprint/33980/ https://www.scopus.com/inward/record.uri?eid=2-s2.0-85113152256&doi=10.1177%2f00375497211038759&partnerID=40&md5=129e7ece1496b7ae0a643243fc3b3eb5 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Agent-based social simulations are typically described in imperative form. While this facilitates implementation as computer programs, it makes implicit the different assumptions made, both about the functional form and the causal ordering involved. As a solution to the problem, a probabilistic graphical model, Action Network (AN), is proposed in this paper for social simulation. Simulation variables are represented by nodes, and causal links by edges. An Action Table is associated with each node, describing incremental probabilistic actions to be performed in response to fuzzy parental states. AN offers a graphical causal model that captures the dynamics of a social process. Details of the formalism are presented along with illustrative examples. Software that implements the formalism is available at http://actionnetwork.epizy.com. © The Author(s) 2021. |
|---|