STPM：一个用于随机过程模型的R软件包。
stpm: an R package for stochastic process model.
作者信息
Zhbannikov Ilya Y, Arbeev Konstantin, Akushevich Igor, Stallard Eric, Yashin Anatoliy I
机构信息
Biodemography of Aging Research Unit (BARU) at Social Science Research Institute, Duke University, 2024 W. Main St., Durham, Box 90420, 27705, NC, USA.
Duke Population Research Institute, Duke University, Durham, Box 90989, 27708-0989, NC, USA.
出版信息
BMC Bioinformatics. 2017 Feb 23;18(1):125. doi: 10.1186/s12859-017-1538-7.
BACKGROUND
The Stochastic Process Model (SPM) represents a general framework for modeling the joint evolution of repeatedly measured variables and time-to-event outcomes observed in longitudinal studies, i.e., SPM relates the stochastic dynamics of variables (e.g., physiological or biological measures) with the probabilities of end points (e.g., death or system failure). SPM is applicable for analyses of longitudinal data in many research areas; however, there are no publicly available software tools that implement this methodology.
RESULTS
We developed an R package stpm for the SPM-methodology. The package estimates several versions of SPM currently available in the literature including discrete- and continuous-time multidimensional models and a one-dimensional model with time-dependent parameters. Also, the package provides tools for simulation and projection of individual trajectories and hazard functions.
CONCLUSION
In this paper, we present the first software implementation of the SPM-methodology by providing an R package stpm, which was verified through extensive simulation and validation studies. Future work includes further improvements of the model. Clinical and academic researchers will benefit from using the presented model and software. The R package stpm is available as open source software from the following links: https://cran.r-project.org/package=stpm (stable version) or https://github.com/izhbannikov/spm (developer version).
背景
随机过程模型（SPM）是一个通用框架，用于对纵向研究中重复测量变量和事件发生时间结局的联合演变进行建模，即SPM将变量的随机动态（例如生理或生物学测量）与终点概率（例如死亡或系统故障）联系起来。SPM适用于许多研究领域的纵向数据分析；然而，目前没有公开可用的软件工具来实现这种方法。
结果
我们为SPM方法开发了一个R包stpm。该包估计了文献中目前可用的几个SPM版本，包括离散时间和连续时间多维模型以及具有时间依赖参数的一维模型。此外，该包还提供了用于模拟和预测个体轨迹及风险函数的工具。
结论
在本文中，我们通过提供一个R包stpm展示了SPM方法的首个软件实现，该包已通过广泛的模拟和验证研究得到验证。未来的工作包括对模型的进一步改进。临床和学术研究人员将从使用所展示的模型和软件中受益。R包stpm可从以下链接作为开源软件获取：https://cran.r-project.org/package=stpm（稳定版本）或https://github.com/izhbannikov/spm（开发者版本）。
Zotero 插件