Transforming Timed Influence Nets into Time Sliced Bayesian Networks
Abstract
The paper presents an algorithm for transforming Timed Influence Nets (TIN) into Time Sliced Bayesian Networks (TSBN). The advantage of TINs lies in their ability to represent both causal and time-sensitive information in a compact and integrated manner. They are used to help a decision maker model the causal and temporal interdependencies among variables in a system. The TIN formalism offers a suite of analysis tools that can be used by a user to analyze the impact of alternate courses of actions on likely outcomes. An even larger, and more robust suite of analysis tools exists for TSBNs. These algorithms also allow analyses that are not available in the TIN formalism, e.g., provision for incorporating real-time information in the form of evidence regarding certain variables and calculating its impact on the rest of the system. The knowledge acquisition process of TSBNs, however, is intractable for large models. This paper is an attempt to combine the advantages of both modeling paradigms, TIN and TSBN, into a single formalism by providing a mapping from a TIN to a TSBN. The proposed formalism uses the TIN approach for the model building and the TSBN for analysis and evaluation. A system analyst, in this combined approach, interacts with a TIN, and the analysis results obtained on the TSBN are mapped back to the TIN, making the transformation completely hidden to the analyst.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2004
- Accession Number
- ADA465981
Entities
People
- Abbas K. Zaidi
- Sajjad Haider
Organizations
- George Mason University