Khanh N. Dinh, Zijin Xiang, Zhihan Liu, Simon Tavaré
{"title":"通过随机森林进行近似贝叶斯计算顺序蒙特卡罗","authors":"Khanh N. Dinh, Zijin Xiang, Zhihan Liu, Simon Tavaré","doi":"arxiv-2406.15865","DOIUrl":null,"url":null,"abstract":"Approximate Bayesian Computation (ABC) is a popular inference method when\nlikelihoods are hard to come by. Practical bottlenecks of ABC applications\ninclude selecting statistics that summarize the data without losing too much\ninformation or introducing uncertainty, and choosing distance functions and\ntolerance thresholds that balance accuracy and computational efficiency. Recent\nstudies have shown that ABC methods using random forest (RF) methodology\nperform well while circumventing many of ABC's drawbacks. However, RF\nconstruction is computationally expensive for large numbers of trees and model\nsimulations, and there can be high uncertainty in the posterior if the prior\ndistribution is uninformative. Here we adapt distributional random forests to\nthe ABC setting, and introduce Approximate Bayesian Computation sequential\nMonte Carlo with random forests (ABC-SMC-(D)RF). This updates the prior\ndistribution iteratively to focus on the most likely regions in the parameter\nspace. We show that ABC-SMC-(D)RF can accurately infer posterior distributions\nfor a wide range of deterministic and stochastic models in different scientific\nareas.","PeriodicalId":501215,"journal":{"name":"arXiv - STAT - Computation","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Approximate Bayesian Computation sequential Monte Carlo via random forests\",\"authors\":\"Khanh N. Dinh, Zijin Xiang, Zhihan Liu, Simon Tavaré\",\"doi\":\"arxiv-2406.15865\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Approximate Bayesian Computation (ABC) is a popular inference method when\\nlikelihoods are hard to come by. Practical bottlenecks of ABC applications\\ninclude selecting statistics that summarize the data without losing too much\\ninformation or introducing uncertainty, and choosing distance functions and\\ntolerance thresholds that balance accuracy and computational efficiency. Recent\\nstudies have shown that ABC methods using random forest (RF) methodology\\nperform well while circumventing many of ABC's drawbacks. However, RF\\nconstruction is computationally expensive for large numbers of trees and model\\nsimulations, and there can be high uncertainty in the posterior if the prior\\ndistribution is uninformative. Here we adapt distributional random forests to\\nthe ABC setting, and introduce Approximate Bayesian Computation sequential\\nMonte Carlo with random forests (ABC-SMC-(D)RF). This updates the prior\\ndistribution iteratively to focus on the most likely regions in the parameter\\nspace. We show that ABC-SMC-(D)RF can accurately infer posterior distributions\\nfor a wide range of deterministic and stochastic models in different scientific\\nareas.\",\"PeriodicalId\":501215,\"journal\":{\"name\":\"arXiv - STAT - Computation\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-06-22\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - STAT - Computation\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2406.15865\",\"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 - Computation","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2406.15865","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Approximate Bayesian Computation sequential Monte Carlo via random forests
Approximate Bayesian Computation (ABC) is a popular inference method when
likelihoods are hard to come by. Practical bottlenecks of ABC applications
include selecting statistics that summarize the data without losing too much
information or introducing uncertainty, and choosing distance functions and
tolerance thresholds that balance accuracy and computational efficiency. Recent
studies have shown that ABC methods using random forest (RF) methodology
perform well while circumventing many of ABC's drawbacks. However, RF
construction is computationally expensive for large numbers of trees and model
simulations, and there can be high uncertainty in the posterior if the prior
distribution is uninformative. Here we adapt distributional random forests to
the ABC setting, and introduce Approximate Bayesian Computation sequential
Monte Carlo with random forests (ABC-SMC-(D)RF). This updates the prior
distribution iteratively to focus on the most likely regions in the parameter
space. We show that ABC-SMC-(D)RF can accurately infer posterior distributions
for a wide range of deterministic and stochastic models in different scientific
areas.