High-dimensional nonparametric density estimation via symmetry and shape constraints

High-dimensional nonparametric density estimation via symmetry and shape constraints
复制标题

通过对称性和形状约束的高维非参数密度估计

DOI:
10.1214/20-aos1972
复制
发表时间:
2019
期刊:
The Annals of Statistics
影响因子:
--
通讯作者:
R. Samworth
R. Samworth
中科院分区:
--
文献类型:
--
作者:
Min Xu;R. Samworth

文献摘要

参考文献

被引文献

相似文献

我们解决了高维非参数密度估计的问题,采取类的对数凹密度$\mathbb{R}^p$,并纳入其对称性假设,这有利于可扩展的估计算法,可以减轻灾难的维数。我们的主要对称性假设是密度的超能级集是$K$-位似的(即凸体$K \subseteq \mathbb{R}^p$的标量倍数)。当$K$是已知的,我们证明了$K$-位似对数凹最大似然估计的基础上$n$独立的意见,从这样的密度有一个最坏情况下的风险界,例如,Hellinger损失的平方为O(n^{-4/5}),与p无关。此外,我们表明,估计是自适应的意义上说,如果数据生成密度承认一个特殊的形式,然后可以达到一个近参数率。我们还提供了最坏情况下的风险范围和自适应的情况下,K$是唯一已知的正定变换,它是完全未知的,必须估计非参数。我们的估计算法是快速的,即使当$n$和$p$是在几十万的顺序,我们说明了强大的有限样本性能的模拟数据,我们的方法。
We tackle the problem of high-dimensional nonparametric density estimation by taking the class of log-concave densities on $\mathbb{R}^p$ and incorporating within it symmetry assumptions, which facilitate scalable estimation algorithms and can mitigate the curse of dimensionality. Our main symmetry assumption is that the super-level sets of the density are $K$-homothetic (i.e. scalar multiples of a convex body $K \subseteq \mathbb{R}^p$). When $K$ is known, we prove that the $K$-homothetic log-concave maximum likelihood estimator based on $n$ independent observations from such a density has a worst-case risk bound with respect to, e.g., squared Hellinger loss, of $O(n^{-4/5})$, independent of $p$. Moreover, we show that the estimator is adaptive in the sense that if the data generating density admits a special form, then a nearly parametric rate may be attained. We also provide worst-case and adaptive risk bounds in cases where $K$ is only known up to a positive definite transformation, and where it is completely unknown and must be estimated nonparametrically. Our estimation algorithms are fast even when $n$ and $p$ are on the order of hundreds of thousands, and we illustrate the strong finite-sample performance of our methods on simulated data.
DOI: --
发表时间: 2018
期刊: PMLR
影响因子: --
作者:
Carpenter, Timothy;Diakonikolas, Ilias;Sidiropoulos, Anastasios;Stewart, Alistair
通讯作者: Stewart, Alistair
通过 Ränyi 散度近似和估计 s 凹面密度。
DOI: 10.1214/15-aos1408
发表时间: 2016
影响因子: 4.5
作者:
Han,Qiyang;Wellner,JonA
通讯作者: Wellner,JonA
DOI: 10.17863/cam.48299
发表时间: 2018-12
期刊: The Annals of Statistics
影响因子: --
作者:
Oliver Y. Feng;Adityanand Guntuboyina;Arlene K. H. Kim;R. Samworth
通讯作者: Oliver Y. Feng;Adityanand Guntuboyina;Arlene K. H. Kim;R. Samworth
DOI: 10.1214/18-aos1753
发表时间: 2017-08
期刊: The Annals of Statistics
影响因子: --
作者:
Q. Han;Tengyao Wang;S. Chatterjee;R. Samworth
通讯作者: Q. Han;Tengyao Wang;S. Chatterjee;R. Samworth
DOI: 10.1214/15-aos1394
发表时间: 2016
影响因子: 4.5
作者:
Doss CR;Wellner JA
通讯作者: Wellner JA