arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1902.02517stat.MLcs.LG
收藏

基于仿真器的统计模型的模型选择:核方法

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

NEC Corporation(日本电气公司);RIKEN AIP(理化学研究所先进智能项目中心);National Institute of Advanced Industrial Science and Technology(产业技术综合研究所);University of Tübingen(蒂宾根大学);Max Planck Institute for Intelligent Systems(马克斯·普朗克智能系统研究所);The Institute of Statistical Mathematics(统计数理研究所)

Takafumi Kajihara, Motonobu Kanagawa, Yuuki Nakaguchi, Kanishka Khandelwal, Kenji Fukumiziu

摘要:

提出了一种基于模拟器的统计模型模型选择的新方法。所提出的方法定义了一个混合的候选模型,然后迭代更新这些模型的权重系数,以及在每个模型中的参数同时,这是通过递归地应用贝叶斯规则,使用最近提出的内核递归ABC算法。该方法的实际优点是,即使建模者缺乏关于每个模型中的参数的适当先验知识,也可以使用该方法。我们证明了所提出的方法的有效性与一些实验,包括生态学和流行病学的动力系统的模型选择。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑