Predicting calving dates with Bayesian Networks

In all forecasts we find an element of uncertainty. Therefore, it is of paramount importance that any modeling tool used for forecasts handles uncertainty in an efficient way. One such tool is the Bayesian network. A Bayesian network represents a forecasting (or decision) problem by a directed, acyclic graph in which the nodes represent the variables (or entities) of a domain, and an edge between any two nodes represent a direct, causal dependency between the entities. The uncertainty in the model in represented by conditional probabilities. There are efficient, computational methods based on local computations for solving problems represented by Bayesian networks. During model development, it is very often the case that a model constructor uses a domain expert as a source of knowledge. In order to revise the model, the domain expert will need a certain level of insight in the modeling technique and the syntax of the representation. The intuitive appearance of a Bayesian network al..