Probabilistic reasoning in intelligent systems: networks of plausible inference
Probabilistic reasoning in intelligent systems: networks of plausible inference
Artificial Intelligence
Causality: models, reasoning, and inference
Causality: models, reasoning, and inference
Constructing the Pignistic Probability Function in a Context of Uncertainty
UAI '89 Proceedings of the Fifth Annual Conference on Uncertainty in Artificial Intelligence
Theory refinement of bayesian networks with hidden variables
Theory refinement of bayesian networks with hidden variables
| Hi-index | 0.00 |
Risk analysis and management are important capabilities in intelligent information and knowledge systems. We present a new approach using directed graph based models for risk analysis and management. Our modelling approach is inspired by and builds on the two level approach of the Transferable Belief Model. The credal level for risk analysis and model construction uses beliefs in causal inference relations among the variables within a domain and a pignistic(betting) level for the decision making. The risk model at the credal level can be transformed into a probabilistic model through a pignistic transformation function. This paper focuses on model construction at the credal level. Our modelling approach captures expert knowledge in a formal and iterative fashion based on the Open World Assumption(OWA) in contrast to Bayesian Network based approaches for managing uncertainty associated with risks which assume all the domain knowledge and data have been captured before hand. As a result, our approach does not require complete knowledges and is well suited for modelling risk in dynamic changing environments where information and knowledge is gathered over time as decisions need to be taken. Its performance is related to the quality of the knowledge at hand at any given time.