Files
Abstract
Models for pest management need to incorporate information from many subject areas,
including biology, agronomy and economics. The information sources are unrelated except
by the structural assumptions of the bio-economic model that is used. Rigorous model
estimation is often at the expense of ignoring much of the available information. This
paper illustrates how a Bayesian approach using the Gibbs sampler can allow the
incorporation of potentially valuable prior information. This approach provides posterior
distributions for the important model relationships, giving an indication of confidence in
the model and providing a natural starting point for stochastic decision making.