Logistic回归模型的统计诊断

统计诊断的主要任务就是通过诊断统计量检测已知观测数据在用既定模型拟合时的合理性,主要是找出数据当中的异常点或强影响点。本文主要研究Logostic回归模型的诊断统计量和诊断统计图。用牛顿迭代法给出Logistic回归模型的极大似然估计值,根据扰动模型得到传统的诊断统计量,结合残差、杠杆值和系数变化三者构造新的诊断统计量,绘制新的诊断统计图,通过模拟研究说明新的诊断统计量的有效性,最后用一个实际案例说明新的诊断方法的应用并进一步验证其优越性。     

One-Step Estimator Paths for Concave Regularization

The statistics literature of the past 15 years has established many favorable properties for sparse diminishing-bias regularization: techniques that can roughly be understood as providing estimation under penalty functions spanning the range of concavity between ?₀ and ?₁ norms. However, lasso ?₁-regularized estimation remains the standard tool for industrial Big Data applications because of its minimal computational cost and the presence of easy-to-apply rules for penalty selection. In response, this article proposes a simple new algorithm framework that requires no more computation than a lasso path: the path of one-step estimators (POSE) does ?₁ penalized regression estimation on a grid of decreasing penalties, but adapts coefficient-specific weights to decrease as a function of the coefficient estimated in the previous path step. This provides sparse diminishing-bias regularization at no extra cost over the fastest lasso algorithms. Moreover, our gamma lasso implementation of POSE is accompanied by a reliable heuristic for the fit degrees of freedom, so that standard information criteria can be applied in penalty selection. We also provide novel results on the distance between weighted-?₁ and ?₀ penalized predictors; this allows us to build intuition about POSE and other diminishing-bias regularization schemes. The methods and results are illustrated in extensive simulations and in application of logistic regression to evaluating the performance of hockey players. Supplementary materials for this article are available online.     

An integrated strategy for the geographical origin traceability of Goji berries by antioxidants characteristic fingerprint based online ultra-performance liquid chromatography-2,2-diphenyl-1-picrylhydrazyl- photodiode array detector-mass spectrometry combined with multivariate statistics analysis

Goji berries are now becoming increasingly popular in the human diet due to their potential health benefits. Unscrupulous traders deliberately mislabel with certain origins to gain illegal profits, which seriously affected the consumers’ benefits. In this study, an online ultra-performance liquid chromatography-2,2-diphenyl-1 -picrylhydrazyl-photodiode array detector-electrospray ionization-quadrupole time of flight mass was developed for rapid screening and identification of the antioxidants from Goji berry; then, the antioxidants characteristic fingerprint was established and explored in the origins discrimination of Goji berries from China combined with multivariate statistics analysis. As a result, twenty-eight compounds were screened from Goji berry extract, 19 of which were identified by accurate molecular and ultraviolet information according to references. Principal components analysis and partial least squares discrimination analysis achieved the accurate classification from the four regions, eight compounds were selected as origin-related antioxidant markers with variable importance in projection >1 and one-way analysis of variance (P<0.05), including rutin, rutin di-hexose, P-coumaric acid tri-hexose, dicaffeoylquinic acid isomer, Quercetin-rhamno-di-hexoside, peak14, peak16, and peak27. This study provides a feasible strategy for the geographical origins discrimination of Goji berries based on antioxidant ingredients difference and will be helpful for improving the quality control level of Goji berries.     

Quality Evaluation of Light- and Dark-Colored Hungarian Honeys,Focusing on Botanical Origin,Antioxidant Capacity and Mineral Content

Melissopalynology, antioxidant capacity and mineral and toxic element contents were analyzed in eight types of Hungarian honeys. Based on color, two groups were distinguished: light honeys comprised acacia, amorpha, phacelia and linden honeys; while dark honeys included sunflower, chestnut, fennel and sage honeys, with 100 to 300 and 700 to 1500 mAU, respectively. The unifloral origin of each sample was supported using pollen analysis. The absorbance of honey correlated positively with antioxidant capacity determined by three different methods (TRC, DPPH, ORAC), and also with mineral content. The exception was the light amber linden honey with significantly higher K content and antiradical activity than other light honeys. The Mn, Zn and Fe contents were the highest in chestnut, sunflower and fennel honeys, respectively. The black meadow sage honey performed best regarding the content of other elements and antioxidant activity. The concentrations of several toxic elements were below the detection limit in the samples, indicating their good quality. The principal component analysis (PCA) revealed correlations between different antioxidant assays and minerals, and furthermore, confirmed the botanical authentication of the honeys based on the studied parameters. To our best knowledge, the present study is the first to provide a complex analysis of quality parameters of eight unifloral Hungarian honeys.     

数理统计在地球化学样品分析质量控制中的应用

使用Excel2013和Minitab两种常用软件对地球化学样品中CaO含量分布情况进行了研究,运用描述性统计、正态分布、背景值与异常方法对实验室分析的广西某地地球化学样品的分析数据进行了质量评价,结果表明:综合运用计算机软件和数理统计方法,能快速找到分析数据的详细信息和数据特征,判别地球化学样品实验室分析数据的准确性,找出元素的背景值和异常值并剔除异常值,比较发现,样品中CaO含量分析数据对数转换后比原始数据的分布更趋于正态分布,相比于X射线荧光光谱方法,用电感耦合等离子体发射光谱法测定的结果更符合正态分布。方法对地球化学调查样品的分析数据质量评价作了有益的尝试,揭示了数理统计方法是地球化学样品分析质量控制的有效手段。     

Tutorial review on validation of liquid chromatography–mass spectrometry methods: Part I

This is the part I of a tutorial review intending to give an overview of the state of the art of method validation in liquid chromatography mass spectrometry (LC–MS) and discuss specific issues that arise with MS (and MS/MS) detection in LC (as opposed to the “conventional” detectors). The Part I briefly introduces the principles of operation of LC–MS (emphasizing the aspects important from the validation point of view, in particular the ionization process and ionization suppression/enhancement); reviews the main validation guideline documents and discusses in detail the following performance parameters: selectivity/specificity/identity, ruggedness/robustness, limit of detection, limit of quantification, decision limit and detection capability. With every method performance characteristic its essence and terminology are addressed, the current status of treating it is reviewed and recommendations are given, how to determine it, specifically in the case of LC–MS methods.     

Two‐dimensional reactive scattering with transmitted quantum trajectories

A simple and easy‐to‐implement method is presented for the study of time‐dependent reaction dynamics by propagating an ensemble of transmitted quantum trajectories. During the trajectory evolution, reflected trajectories are gradually removed and all the remaining trajectories represent the transmitted subensemble. The removal process of reflected trajectories avoids numerical instabilities arising from node formation in the reactant region, and allows stable long‐time propagation of transmitted trajectories. This method is applied to a two‐dimensional model chemical reaction. Excellent computational results are obtained for the time‐dependent reaction probabilities evaluated by the time integration of the probability flux. © 2014 Wiley Periodicals, Inc.     

Modeling of dynamics, heat transfer, and combustion in two-phase turbulent flows: 2. Flows with heat transfer and combustion

The objective of this part of the paper is to summarize the information concerning the authors' works in the field of simulation of two-phase gas-particle turbulent flows with heat transfer and combustion. A kinetic equation had been derived for the probability density function (PDF) of the particle velocity, temperature, and mass distributions in turbulent flows. This PDF equation is used for the construction of the governing conservation equations of mass, momentum, and heat transfer in the dispersed particle phase.The numerical scheme incorporates two-phase fluid dynamics, convective and radiative heat transfer, and combustion. The proposed models have been applied to the calculation of various particle-laden turbulent flows in jets, combustion and gasification chambers, and furnaces.     

玻璃态高聚物细观损伤断裂统计力学

对玻璃态高聚物内部细观损伤断裂判据、机理、动力学及统计模型的最新进展，进行了较为全面、系统的介绍和总结，并简要概述了细观损伤统计描述在金属材料及玻璃态高聚物领域里的应用。     

多模式自适应重要抽样法及其应用

针对多模式的可靠性分析，研究了其失效概率计算的自适应重要抽样法，该方法用模拟退火 算法来自动调整每个失效模式的重要抽样函数，使其逐渐趋近于估计方差最小的重要抽样 函数. 对于多个模式系统失效概率的计算，采用混合加权自适应重要抽样的方法, 反映了每个 失效模式对系统失效概率的贡献；对于系统失效模式所含基本变量不全相同的情况，提出了 扩展自适应重要抽样法, 来统一所有失效模式中的基本变量，从而使得混合自适应 重要抽样, 可以方便地求解变量不全相同时的系统失效概率. 对估计值方差和变异系数的计算公 式进行了推导. 验证算例结果, 充分说明方法的合理性与可行性.