Bayesian estimation in forest surveys when samples or prior information are fuzzy

Fuzzy phenomena are frequently encountered in forestry due to the complexity of the systems under investigation. In this paper, existing inference methods are presented to accommodate situations in which random events take place in fuzzy circumstances. Fuzzy concepts are incorporated into these commonly accepted approaches to account for uncertainties caused by factors other than randomness which traditional statistical approaches deal primarily with. Based on two extensions of likelihood functions for fuzzy samples, Bayesian estimates are generalized for use when sample information and prior distribution of parameters are fuzzy. The estimators are developed for some typical membership functions for fuzzy samples and fuzzy priors. Applications are illustrated using examples from forestry.