arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~

CosmoFit:用于用户自定义H(z)模型的贝叶斯宇宙学参数估计的图形用户界面

CosmoFit: A Graphical User Interface for Bayesian Cosmological Parameter Estimation with User-Defined H(z) Models

Giovany Cruz

arXiv 2608.23588首次发表:更新:

AI 中文总结

CosmoFit是一款桌面图形界面,可让用户自定义H(z)模型,通过Cobaya和GetDist完成贝叶斯宇宙学参数估计,经测试其结果与直接用Cobaya的一致性在1%内,易用且可复现。

AI 中文摘要

我们提出了CosmoFit,这是一款用于背景宇宙学贝叶斯参数推断的桌面图形界面。该应用旨在简化宇宙学分析,减少手动编程每一步的需求,同时保持结果的透明性和可重复性。用户可以定义膨胀历史H(z),指定固定参数和待采样参数,分配先验范围和提议尺度,选择可用的观测似然函数,并从统一的图形界面运行马尔可夫链蒙特卡洛分析。CosmoFit使用Cobaya进行贝叶斯采样,使用GetDist进行后验分析和可视化。当前版本支持宇宙计时器数据,以及通过Cobaya提供的官方Pantheon+、Pantheon+SH0ES和Union3超新星似然函数。对于每次运行,应用会存储配置文件、元数据、执行日志、链以及检查和复现分析所需的数值摘要。通过单元测试、使用合成数据的参数恢复测试,以及与直接在Cobaya中进行的分析的独立比较,对CosmoFit的性能进行了评估。两种方法获得的后验均值的一致性在1%以内。此外,合成数据测试在预期的可信区域内恢复了基准参数。作为说明性示例,我们使用Pantheon+SH0ES进行了空间平坦的wCDM分析,并直接从界面获得了边缘化约束和三角图。在其当前范围内,CosmoFit为涉及背景宇宙膨胀显式模型的研究和学术应用提供了一个易用、透明且可重复的工具。

英文摘要

We present CosmoFit, a desktop graphical interface for Bayesian parameter inference in background cosmology. The application is designed to simplify cosmological analyses by reducing the need to manually program each step, while maintaining the transparency and reproducibility of the results. Users can define an expansion history H(z), specify fixed and sampled parameters, assign prior ranges and proposal scales, select the available observational likelihoods, and run Markov chain Monte Carlo analyses from a unified graphical interface. CosmoFit uses Cobaya for Bayesian sampling and GetDist for posterior analysis and visualization. The current version supports cosmic-chronometer data and the official Pantheon+, Pantheon+SH0ES, and Union3 supernova likelihoods available through Cobaya. For each run, the application stores the configuration files, metadata, execution logs, chains, and numerical summaries required to inspect and reproduce the analysis. The performance of CosmoFit was evaluated through unit tests, a parameter-recovery test using synthetic data, and an independent comparison with an analysis performed directly in Cobaya. The posterior mean values obtained with both procedures agree to within less than one percent. In addition, the synthetic-data test recovers the fiducial parameters within the expected credible regions. As an illustrative example, we perform a spatially flat wCDM analysis using Pantheon+SH0ES and obtain marginalized constraints and triangle plots directly from the interface. Within its current scope, CosmoFit provides an accessible, transparent, and reproducible tool for research and academic applications involving explicit models of the background cosmological expansion.

论文原文

arXiv 摘要页 · PDF 原文 · HTML 原文

↑