arXiv - STAT - Methodology最新文献

Poisson approximate likelihood compared to the particle filter 泊松近似似然与粒子滤波的比较

arXiv - STAT - Methodology Pub Date : 2024-09-18 DOI: arxiv-2409.12173

Yize Hao, Aaron A. Abkemeier, Edward L. Ionides

{"title":"Poisson approximate likelihood compared to the particle filter","authors":"Yize Hao, Aaron A. Abkemeier, Edward L. Ionides","doi":"arxiv-2409.12173","DOIUrl":"https://doi.org/arxiv-2409.12173","url":null,"abstract":"Filtering algorithms are fundamental for inference on partially observed\u0000stochastic dynamic systems, since they provide access to the likelihood\u0000function and hence enable likelihood-based or Bayesian inference. A novel\u0000Poisson approximate likelihood (PAL) filter was introduced by Whitehouse et al.\u0000(2023). PAL employs a Poisson approximation to conditional densities, offering\u0000a fast approximation to the likelihood function for a certain subset of\u0000partially observed Markov process models. A central piece of evidence for PAL\u0000is the comparison in Table 1 of Whitehouse et al. (2023), which claims a large\u0000improvement for PAL over a standard particle filter algorithm. This evidence,\u0000based on a model and data from a previous scientific study by Stocks et al.\u0000(2020), might suggest that researchers confronted with similar models should\u0000use PAL rather than particle filter methods. Taken at face value, this evidence\u0000also reduces the credibility of Stocks et al. (2020) by indicating a\u0000shortcoming with the numerical methods that they used. However, we show that\u0000the comparison of log-likelihood values made by Whitehouse et al. (2023) is\u0000flawed because their PAL calculations were carried out using a dataset scaled\u0000differently from the previous study. If PAL and the particle filter are applied\u0000to the same data, the advantage claimed for PAL disappears. On simulations\u0000where the model is correctly specified, the particle filter outperforms PAL.","PeriodicalId":501425,"journal":{"name":"arXiv - STAT - Methodology","volume":"44 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142256324","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Optimising the Trade-Off Between Type I and Type II Errors: A Review and Extensions 优化 I 型和 II 型误差之间的权衡：回顾与扩展

arXiv - STAT - Methodology Pub Date : 2024-09-18 DOI: arxiv-2409.12081

Andrew P Grieve

{"title":"Optimising the Trade-Off Between Type I and Type II Errors: A Review and Extensions","authors":"Andrew P Grieve","doi":"arxiv-2409.12081","DOIUrl":"https://doi.org/arxiv-2409.12081","url":null,"abstract":"In clinical studies upon which decisions are based there are two types of\u0000errors that can be made: a type I error arises when the decision is taken to\u0000declare a positive outcome when the truth is in fact negative, and a type II\u0000error arises when the decision is taken to declare a negative outcome when the\u0000truth is in fact positive. Commonly the primary analysis of such a study\u0000entails a two-sided hypothesis test with a type I error rate of 5% and the\u0000study is designed to have a sufficiently low type II error rate, for example\u000010% or 20%. These values are arbitrary and often do not reflect the clinical,\u0000or regulatory, context of the study and ignore both the relative costs of\u0000making either type of error and the sponsor's prior belief that the drug is\u0000superior to either placebo, or a standard of care if relevant. This simplistic\u0000approach has recently been challenged by numerous authors both from a\u0000frequentist and Bayesian perspective since when resources are constrained there\u0000will be a need to consider a trade-off between type I and type II errors. In\u0000this paper we review proposals to utilise the trade-off by formally\u0000acknowledging the costs to optimise the choice of error rates for simple, point\u0000null and alternative hypotheses and extend the results to composite, or\u0000interval hypotheses, showing links to the Probability of Success of a clinical\u0000study.","PeriodicalId":501425,"journal":{"name":"arXiv - STAT - Methodology","volume":"123 14 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142256370","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Bias Reduction in Matched Observational Studies with Continuous Treatments: Calipered Non-Bipartite Matching and Bias-Corrected Estimation and Inference 在连续治疗的匹配观察研究中减少偏差：Calipered Non-Bipartite Matching 和偏差校正估计与推论

arXiv - STAT - Methodology Pub Date : 2024-09-18 DOI: arxiv-2409.11701

Anthony Frazier, Siyu Heng, Wen Zhou

{"title":"Bias Reduction in Matched Observational Studies with Continuous Treatments: Calipered Non-Bipartite Matching and Bias-Corrected Estimation and Inference","authors":"Anthony Frazier, Siyu Heng, Wen Zhou","doi":"arxiv-2409.11701","DOIUrl":"https://doi.org/arxiv-2409.11701","url":null,"abstract":"Matching is a commonly used causal inference framework in observational\u0000studies. By pairing individuals with different treatment values but with the\u0000same values of covariates (i.e., exact matching), the sample average treatment\u0000effect (SATE) can be consistently estimated and inferred using the classic\u0000Neyman-type (difference-in-means) estimator and confidence interval. However,\u0000inexact matching typically exists in practice and may cause substantial bias\u0000for the downstream treatment effect estimation and inference. Many methods have\u0000been proposed to reduce bias due to inexact matching in the binary treatment\u0000case. However, to our knowledge, no existing work has systematically\u0000investigated bias due to inexact matching in the continuous treatment case. To\u0000fill this blank, we propose a general framework for reducing bias in inexactly\u0000matched observational studies with continuous treatments. In the matching\u0000stage, we propose a carefully formulated caliper that incorporates the\u0000information of both the paired covariates and treatment doses to better tailor\u0000matching for the downstream SATE estimation and inference. In the estimation\u0000and inference stage, we propose a bias-corrected Neyman estimator paired with\u0000the corresponding bias-corrected variance estimator to leverage the information\u0000on propensity density discrepancies after inexact matching to further reduce\u0000the bias due to inexact matching. We apply our proposed framework to COVID-19\u0000social mobility data to showcase differences between classic and bias-corrected\u0000SATE estimation and inference.","PeriodicalId":501425,"journal":{"name":"arXiv - STAT - Methodology","volume":"17 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142256371","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Forecasting age distribution of life-table death counts via α-transformation 通过α变换预测生命表死亡人数的年龄分布

arXiv - STAT - Methodology Pub Date : 2024-09-18 DOI: arxiv-2409.11658

Han Lin Shang, Steven Haberman

引用次数: 0

E-Values for Exponential Families: the General Case 指数族的 E 值：一般情况

arXiv - STAT - Methodology Pub Date : 2024-09-17 DOI: arxiv-2409.11134

Yunda Hao, Peter Grünwald

引用次数: 0

Cointegrated Matrix Autoregression Models 协整矩阵自回归模型

arXiv - STAT - Methodology Pub Date : 2024-09-17 DOI: arxiv-2409.10860

Zebang Li, Han Xiao

引用次数: 0

Calibrated Multivariate Regression with Localized PIT Mappings 使用本地化 PIT 映射的校准多元回归

arXiv - STAT - Methodology Pub Date : 2024-09-17 DOI: arxiv-2409.10855

Lucas Kock, G. S. Rodrigues, Scott A. Sisson, Nadja Klein, David J. Nott

{"title":"Calibrated Multivariate Regression with Localized PIT Mappings","authors":"Lucas Kock, G. S. Rodrigues, Scott A. Sisson, Nadja Klein, David J. Nott","doi":"arxiv-2409.10855","DOIUrl":"https://doi.org/arxiv-2409.10855","url":null,"abstract":"Calibration ensures that predicted uncertainties align with observed\u0000uncertainties. While there is an extensive literature on recalibration methods\u0000for univariate probabilistic forecasts, work on calibration for multivariate\u0000forecasts is much more limited. This paper introduces a novel post-hoc\u0000recalibration approach that addresses multivariate calibration for potentially\u0000misspecified models. Our method involves constructing local mappings between\u0000vectors of marginal probability integral transform values and the space of\u0000observations, providing a flexible and model free solution applicable to\u0000continuous, discrete, and mixed responses. We present two versions of our\u0000approach: one uses K-nearest neighbors, and the other uses normalizing flows.\u0000Each method has its own strengths in different situations. We demonstrate the\u0000effectiveness of our approach on two real data applications: recalibrating a\u0000deep neural network's currency exchange rate forecast and improving a\u0000regression model for childhood malnutrition in India for which the multivariate\u0000response has both discrete and continuous components.","PeriodicalId":501425,"journal":{"name":"arXiv - STAT - Methodology","volume":"44 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142256377","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Comparison of g-estimation approaches for handling symptomatic medication at multiple timepoints in Alzheimer's Disease with a hypothetical strategy 阿尔茨海默病多时间点症状用药的 g 值估算方法与假设策略的比较

arXiv - STAT - Methodology Pub Date : 2024-09-17 DOI: arxiv-2409.10943

Florian Lasch, Lorenzo Guizzaro, Wen Wei Loh

{"title":"Comparison of g-estimation approaches for handling symptomatic medication at multiple timepoints in Alzheimer's Disease with a hypothetical strategy","authors":"Florian Lasch, Lorenzo Guizzaro, Wen Wei Loh","doi":"arxiv-2409.10943","DOIUrl":"https://doi.org/arxiv-2409.10943","url":null,"abstract":"For handling intercurrent events in clinical trials, one of the strategies\u0000outlined in the ICH E9(R1) addendum targets the hypothetical scenario of\u0000non-occurrence of the intercurrent event. While this strategy is often\u0000implemented by setting data after the intercurrent event to missing even if\u0000they have been collected, g-estimation allows for a more efficient estimation\u0000by using the information contained in post-IE data. As the g-estimation methods\u0000have largely developed outside of randomised clinical trials, optimisations for\u0000the application in clinical trials are possible. In this work, we describe and\u0000investigate the performance of modifications to the established g-estimation\u0000methods, leveraging the assumption that some intercurrent events are expected\u0000to have the same impact on the outcome regardless of the timing of their\u0000occurrence. In a simulation study in Alzheimer disease, the modifications show\u0000a substantial efficiency advantage for the estimation of an estimand that\u0000applies the hypothetical strategy to the use of symptomatic treatment while\u0000retaining unbiasedness and adequate type I error control.","PeriodicalId":501425,"journal":{"name":"arXiv - STAT - Methodology","volume":"13 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142256376","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Data-driven stochastic 3D modeling of the nanoporous binder-conductive additive phase in battery cathodes 电池阴极中纳米多孔粘合剂导电添加剂相的数据驱动随机三维建模

arXiv - STAT - Methodology Pub Date : 2024-09-17 DOI: arxiv-2409.11080

Phillip Gräfensteiner, Markus Osenberg, André Hilger, Nicole Bohn, Joachim R. Binder, Ingo Manke, Volker Schmidt, Matthias Neumann

{"title":"Data-driven stochastic 3D modeling of the nanoporous binder-conductive additive phase in battery cathodes","authors":"Phillip Gräfensteiner, Markus Osenberg, André Hilger, Nicole Bohn, Joachim R. Binder, Ingo Manke, Volker Schmidt, Matthias Neumann","doi":"arxiv-2409.11080","DOIUrl":"https://doi.org/arxiv-2409.11080","url":null,"abstract":"A stochastic 3D modeling approach for the nanoporous binder-conductive\u0000additive phase in hierarchically structured cathodes of lithium-ion batteries\u0000is presented. The binder-conductive additive phase of these electrodes consists\u0000of carbon black, polyvinylidene difluoride binder and graphite particles. For\u0000its stochastic 3D modeling, a three-step procedure based on methods from\u0000stochastic geometry is used. First, the graphite particles are described by a\u0000Boolean model with ellipsoidal grains. Second, the mixture of carbon black and\u0000binder is modeled by an excursion set of a Gaussian random field in the\u0000complement of the graphite particles. Third, large pore regions within the\u0000mixture of carbon black and binder are described by a Boolean model with\u0000spherical grains. The model parameters are calibrated to 3D image data of\u0000cathodes in lithium-ion batteries acquired by focused ion beam scanning\u0000electron microscopy. Subsequently, model validation is performed by comparing\u0000model realizations with measured image data in terms of various morphological\u0000descriptors that are not used for model fitting. Finally, we use the stochastic\u00003D model for predictive simulations, where we generate virtual, yet realistic,\u0000image data of nanoporous binder-conductive additives with varying amounts of\u0000graphite particles. Based on these virtual nanostructures, we can investigate\u0000structure-property relationships. In particular, we quantitatively study the\u0000influence of graphite particles on effective transport properties in the\u0000nanoporous binder-conductive additive phase, which have a crucial impact on\u0000electrochemical processes in the cathode and thus on the performance of battery\u0000cells.","PeriodicalId":501425,"journal":{"name":"arXiv - STAT - Methodology","volume":"187 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142256384","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Chasing Shadows: How Implausible Assumptions Skew Our Understanding of Causal Estimands 追逐阴影：匪夷所思的假设如何歪曲我们对因果估计的理解

arXiv - STAT - Methodology Pub Date : 2024-09-17 DOI: arxiv-2409.11162

Stijn Vansteelandt, Kelly Van Lancker

引用次数: 0