كاهش خطاي مدل سازي به كمك ميانگين گيري بيزي

Decreasing Modeling Errors using Bayesian Model Averaging

. In statistical modeling, usually, there is no model that has a completely desirable fitting to the data. Often there is a class of models with reasonable fitness. Tra¬ditional modeling methods, based on some fitting criteria, select a goal model as the best. This modeling approach dose not takes into account the really uncertainty of modeling procedure and increases the errors. Model averaging is a powerful modeling approach in which a weighted average of all models or a subset of them is considered as the desired model. This method increases the predictive efficiency and reduces the modeling errors.