Multimodel Inference based on Smoothed Information Criteria

Multimodel inference makes statistical inferences from a set of plausible models rather than from a single model. In this paper, we focus on the multimodel inference based on smoothed information criteria proposed by monographs Buckland et al. (1997) and Burnham & Anderson (2003), which are termed as smoothed AIC (SAIC) and smoothed BIC (SBIC) methods. Because of simplicity and applicability, these methods are very widely used in many fields. By using an illustrative example and deriving limiting properties for the weights, we find that the existing variance estimation for SAIC is not proper, but for SBIC, it is. We propose a simulation-based inference for SAIC. The simulation results show its promising performance