{"title":"贝叶斯非参数准似然法","authors":"Antonio R. Linero","doi":"arxiv-2405.20601","DOIUrl":null,"url":null,"abstract":"A recent trend in Bayesian research has been revisiting generalizations of\nthe likelihood that enable Bayesian inference without requiring the\nspecification of a model for the data generating mechanism. This paper focuses\non a Bayesian nonparametric extension of Wedderburn's quasi-likelihood, using\nBayesian additive regression trees to model the mean function. Here, the\nanalyst posits only a structural relationship between the mean and variance of\nthe outcome. We show that this approach provides a unified, computationally\nefficient, framework for extending Bayesian decision tree ensembles to many new\nsettings, including simplex-valued and heavily heteroskedastic data. We also\nintroduce Bayesian strategies for inferring the dispersion parameter of the\nquasi-likelihood, a task which is complicated by the fact that the\nquasi-likelihood itself does not contain information about this parameter;\ndespite these challenges, we are able to inject updates for the dispersion\nparameter into a Markov chain Monte Carlo inference scheme in a way that, in\nthe parametric setting, leads to a Bernstein-von Mises result for the\nstationary distribution of the resulting Markov chain. We illustrate the\nutility of our approach on a variety of both synthetic and non-synthetic\ndatasets.","PeriodicalId":501323,"journal":{"name":"arXiv - STAT - Other Statistics","volume":"41 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Bayesian Nonparametric Quasi Likelihood\",\"authors\":\"Antonio R. Linero\",\"doi\":\"arxiv-2405.20601\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"A recent trend in Bayesian research has been revisiting generalizations of\\nthe likelihood that enable Bayesian inference without requiring the\\nspecification of a model for the data generating mechanism. This paper focuses\\non a Bayesian nonparametric extension of Wedderburn's quasi-likelihood, using\\nBayesian additive regression trees to model the mean function. Here, the\\nanalyst posits only a structural relationship between the mean and variance of\\nthe outcome. We show that this approach provides a unified, computationally\\nefficient, framework for extending Bayesian decision tree ensembles to many new\\nsettings, including simplex-valued and heavily heteroskedastic data. We also\\nintroduce Bayesian strategies for inferring the dispersion parameter of the\\nquasi-likelihood, a task which is complicated by the fact that the\\nquasi-likelihood itself does not contain information about this parameter;\\ndespite these challenges, we are able to inject updates for the dispersion\\nparameter into a Markov chain Monte Carlo inference scheme in a way that, in\\nthe parametric setting, leads to a Bernstein-von Mises result for the\\nstationary distribution of the resulting Markov chain. We illustrate the\\nutility of our approach on a variety of both synthetic and non-synthetic\\ndatasets.\",\"PeriodicalId\":501323,\"journal\":{\"name\":\"arXiv - STAT - Other Statistics\",\"volume\":\"41 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-05-31\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - STAT - Other Statistics\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2405.20601\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - STAT - Other Statistics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2405.20601","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
A recent trend in Bayesian research has been revisiting generalizations of
the likelihood that enable Bayesian inference without requiring the
specification of a model for the data generating mechanism. This paper focuses
on a Bayesian nonparametric extension of Wedderburn's quasi-likelihood, using
Bayesian additive regression trees to model the mean function. Here, the
analyst posits only a structural relationship between the mean and variance of
the outcome. We show that this approach provides a unified, computationally
efficient, framework for extending Bayesian decision tree ensembles to many new
settings, including simplex-valued and heavily heteroskedastic data. We also
introduce Bayesian strategies for inferring the dispersion parameter of the
quasi-likelihood, a task which is complicated by the fact that the
quasi-likelihood itself does not contain information about this parameter;
despite these challenges, we are able to inject updates for the dispersion
parameter into a Markov chain Monte Carlo inference scheme in a way that, in
the parametric setting, leads to a Bernstein-von Mises result for the
stationary distribution of the resulting Markov chain. We illustrate the
utility of our approach on a variety of both synthetic and non-synthetic
datasets.