Model Selection for Simulator-based Statistical Models: A Kernel Approach
This addresses model selection challenges for researchers in fields like ecology and epidemiology when prior knowledge is limited, though it appears incremental as it builds on existing kernel recursive ABC methods.
The paper tackles model selection for simulator-based statistical models by proposing a kernel approach that iteratively updates model weights and parameters using Bayes' rule, demonstrating effectiveness in experiments with dynamical systems in ecology and epidemiology.
We propose a novel approach to model selection for simulator-based statistical models. The proposed approach defines a mixture of candidate models, and then iteratively updates the weight coefficients for those models as well as the parameters in each model simultaneously; this is done by recursively applying Bayes' rule, using the recently proposed kernel recursive ABC algorithm. The practical advantage of the method is that it can be used even when a modeler lacks appropriate prior knowledge about the parameters in each model. We demonstrate the effectiveness of the proposed approach with a number of experiments, including model selection for dynamical systems in ecology and epidemiology.