This is a decommissioned version of ERA which is running to enable completion of migration processes. All new collections and items and all edits to existing items should go to our new ERA instance at https://ualberta.scholaris.ca - Please contact us at erahelp@ualberta.ca for assistance!
- 388 views
- 612 downloads
Structure Learning of Causal Bayesian Networks: A Survey
-
- Author(s) / Creator(s)
-
Technical report TR11-01. Causality is a fundamental concept in reasoning. The effectiveness of many reasoning tasks depends on the understanding of the underlying cause-effect relationships. Therefore, the notion of causality has been explored in a wide range of disciplines. Causal discovery, however, was not modeled as a machine learning task until recently. Many learning approaches have recently been developed and applied to capture causation. The most frequently used approach among them is learning causal Bayesian networks (CBNs). A powerful calculus, capable of causal reasoning, has been formalized through CBNs. In this paper, we reviewed the fundamentals of learning causal structures using CBNs. We distinguished between observation and intervention, a crucial concept for learning CBNs. We reviewed some methods for learning from observational and interventional data. We have noted that, as a growing field of research, learning CBN structure is being investigated with increasingly difficult problems and possibilities are arising for incorporating it to other learning problems, such as active learning. | TRID-ID TR11-01
-
- Date created
- 2011
-
- Subjects / Keywords
-
- Type of Item
- Report
-
- License
- Attribution 3.0 International