Bayesian Inference for the Weights in Logarithmic Pooling

Combining distributions is an important issue in decision theory and Bayesian inference. Logarithmic pooling a popular method to aggregate expert opinions by using set of weights that reflect the reliability each information source. However, resulting pooled distribution depends heavily on given opinion/prior thus careful consideration must be choice weights. In this paper we review extend statistical logarithmic pooling, focusing assignment hierarchical prior distribution. We explore several applications, such as estimation survival probabilities, meta-analysis melding deterministic models population growth epidemics. show it possible learn from data, although identifiability issues may arise for some configurations priors data. Furthermore, how approach leads posterior are able accommodate prior-data conflict complex models.