# Model Selection for Simulator-based Statistical Models: A Kernel Approach

## TL;DR

This paper introduces a kernel-based iterative method for model selection in simulator-based statistical models, capable of handling situations with limited prior knowledge, demonstrated through ecological and epidemiological examples.

## Contribution

It presents a novel mixture model approach combined with recursive Bayesian updating using kernel ABC, enabling effective model selection without requiring detailed prior information.

## Key findings

- Effective in ecological and epidemiological models
- Handles limited prior knowledge
- Demonstrates improved model selection accuracy

## Abstract

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.