arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2004.06857stat.CO
收藏

多元Poisson-lognormal分布族用于多元计数数据聚类

A parsimonious family of multivariate Poisson-lognormal distributions for clustering multivariate count data

Sanjeena Subedi, Ryan Browne

摘要:

多变量计数数据通常通过生物信息学、文本挖掘或体育分析中的高通量测序技术遇到。虽然泊松分布似乎很适合这些计数数据,但其多变量扩展在计算上是昂贵的。在大多数情况下,假设变量之间相互独立,但这没有考虑到数据中通常观察到的变量之间的相关性。最近,多元Poisson-lognormal(MPLN)模型的混合已被用于分析具有相关结构的此类多元计数测量。在MPLN模型中,使用以潜在多变量高斯变量为条件的独立泊松分布对每个计数进行建模。由于这种分层结构,MPLN模型可以解释过度分散,而不是传统的泊松分布,并允许变量之间的相关性。不是依赖于计算效率低的基于蒙特卡罗的估计框架,而是在此使用基于快速变分EM的框架来进行参数估计。此外,提出了一个简约的家庭的混合泊松对数正态分布的分解的协方差阵和这些分解施加约束。使用模拟和基准数据集,这些模型的效用。

英文摘要:

Multivariate count data are commonly encountered through high-throughput sequencing technologies in bioinformatics, text mining, or in sports analytics. Although the Poisson distribution seems a natural fit to these count data, its multivariate extension is computationally expensive.In most cases mutual independence among the variables is assumed, however this fails to take into account the correlation among the variables usually observed in the data. Recently, mixtures of multivariate Poisson-lognormal (MPLN) models have been used to analyze such multivariate count measurements with a dependence structure. In the MPLN model, each count is modeled using an independent Poisson distribution conditional on a latent multivariate Gaussian variable. Due to this hierarchical structure, the MPLN model can account for over-dispersion as opposed to the traditional Poisson distribution and allows for correlation between the variables. Rather than relying on a Monte Carlo-based estimation framework which is computationally inefficient, a fast variational-EM based framework is used here for parameter estimation. Further, a parsimonious family of mixtures of Poisson-lognormal distributions are proposed by decomposing the covariance matrix and imposing constraints on these decompositions. Utility of such models is shown using simulated and benchmark datasets.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑