稀疏性下分类的错误发现率阈值
On false discovery rate thresholding for classification under sparsity
摘要:
我们研究了错误发现率(FDR)阈值的性质,作为一个分类过程。假设“0”类(空)具有已知密度,而“1”类(备选)通过平移或缩放从“0”类获得。此外,假设“1”类具有少量元素w.r.t.“0”类(稀疏性)。我们专注于Subbotin家族的密度,包括高斯和拉普拉斯模型。对于FDR门限的超额风险,导出了非渐近预言不等式.这些不等式导致超额风险收敛到零的显式速率,因为要分类的项目的数量m趋于无穷大,并且在贝叶斯规则的功率远离0和1的情况下。此外,这些理论研究表明,作为m的函数,FDR阈值化的目标水平$α_m$的明确选择。我们的预言不等式理论上表明,所得到的FDR阈值适应于数据中包含的未知稀疏性制度。数值实验说明了这一特性。
英文摘要:
We study the properties of false discovery rate (FDR) thresholding, viewed as a classification procedure. The "0"-class (null) is assumed to have a known density while the "1"-class (alternative) is obtained from the "0"-class either by translation or by scaling. Furthermore, the "1"-class is assumed to have a small number of elements w.r.t. the "0"-class (sparsity). We focus on densities of the Subbotin family, including Gaussian and Laplace models. Nonasymptotic oracle inequalities are derived for the excess risk of FDR thresholding. These inequalities lead to explicit rates of convergence of the excess risk to zero, as the number m of items to be classified tends to infinity and in a regime where the power of the Bayes rule is away from 0 and 1. Moreover, these theoretical investigations suggest an explicit choice for the target level $α_m$ of FDR thresholding, as a function of m. Our oracle inequalities show theoretically that the resulting FDR thresholding adapts to the unknown sparsity regime contained in the data. This property is illustrated with numerical experiments.
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。