An EM-type algorithm for multivariate mixture models

This paper introduces a new approach, based on dependent univariate GLMs, for fitting multivariate mixture models. This approach is a multivariate generalization of the method for univariate mixtures presented by Hinde (1982). Its accuracy and efficiency are compared with direct maximization of the log-likelihood. Using a simulation study, we also compare the efficiency of Monte Carlo and Gaussian quadrature methods for approximating the mixture distribution. The new approach with Gaussian quadrature outperforms the alternative methods considered. The work is motivated by the multivariate mixture models which have been proposed for modelling changes of employment states at an individual level. Similar formulations are of interest for modelling movement between other social and economic states and multivariate mixture models also occur in biostatistics and epidemiology.     

氧化镝、氟化钠分解电压的测定及其热力学分析

用线性扫描伏安法测得氧化镝和氟化钠的分解电压(以玻璃化碳为研究电极)分别是 E~(900℃)=1.34V 和 E~(970℃)=2.71V。热力学分析表明实测分解电压与理论分解电压相差较大的主要原因是阳极上的去极化作用——阳极上形成四氟化碳(对 NaF)或一氧化碳(对Dy_2O_3)。     

农户化肥减量施用意向影响因素及其效应分解——基于VBN-TPB的实证分析

根据价值-信念-规范理论和计划行为理论,基于粮食主产区山东省549户农户调研数据,采用结构方程模型分析农户耕地化肥减量施用行为意向影响因素及其效应分解。结果表明：农户化肥减量施用价值感知显著正向影响其行为态度、主观规范和知觉行为控制,显著负向影响其责任归属;行为态度、主观规范、知觉行为控制和责任归属均显著促进行为意向,其中知觉行为控制促进作用最大。效应分解上,主观规范对行为意向的直接正效应大于其通过责任归属的间接正效应;价值感知对责任归属的直接负效应大于其通过主观规范的间接正效应。多群组分析结果表明,兼业程度和耕地规模调节变量在不同假设路径中的影响存在较大差异。随着兼业程度的增加,价值感知对行为态度、主观规范和知觉行为控制的正向作用均大致呈逐渐上升态势,而随着耕地经营规模的扩大,价值感知的正向作用则呈减小态势。对化肥减量施用行为意向的影响方面,兼业程度分组中,责任归属对纯农户的影响最大、主观规范对III兼农户的影响最大、知觉行为控制对II兼农户的影响最大;耕地规模分组中,责任归属和知觉行为控制对中等规模农户的影响最大。     

基于二次分解策略和模糊时间序列模型的航空客运需求预测研究

提高航空客运需求预测的准确性对于航空公司以及整个航空运输系统的发展都具有重要的现实意义。以往研究普遍采用单一分解策略去处理航空客运需求时序中存在的复杂特征,以此提升组合模型的预测性能。然而传统的分解策略存在着特征提取不完全、分解方法带有固有缺陷等问题,导致组合模型预测效果不能得到充分的提升。为此,本文提出一种基于二次分解策略和模糊时间序列模型的航空客运需求预测方法。该方法首先利用季节调整模型（X12-ARIMA）将原始时序分解成季节成分序列与季节调整后序列,继而利用改进的自适应噪声集成经验模态分解方法（ICEEMDAN）将季节调整后序列分解成一系列不同时间尺度的本征模态函数（IMF）和残差序列（Residue）。然后使用基于模糊C均值算法（FCM）划分论域区间的FTS模型对季节成分序列、各IMF分量以及残差序列分别进行预测。最后将各分量序列的预测结果进行集成,重构出航空客运需求的预测值。实证结果表明,本文所提出的二次分解策略表现显著优于传统的分解策略,并且本文所提出模型对于航空客运需求预测有着较高的准确性。     

An improved algorithm for reliability bounds of multistate networks

Abstract

Indirect approaches based on minimal path vectors (d-MPs) and/or minimal cut vectors (d-MCs) are reported to be efficient for the reliability evaluation of multistate networks. Given the need to find more efficient evaluation methods for exact reliability, such techniques may still be cumbersome when the size of the network and the states of component are relatively large. Alternatively, computing reliability bounds can provide approximated reliability with less computational effort. Based on Bai’s exact and indirect reliability evaluation algorithm, an improved algorithm is proposed in this study, which provides sequences of upper and lower reliability bounds of multistate networks. Novel heuristic rules with a pre-specified value to filter less important sets of unspecified states are then developed and incorporated into the algorithm. Computational experiments comparing the proposed methods with an existing direct bounding algorithm show that the new algorithms can provide tight reliability bounds with less computational effort, especially for the proposed algorithm with heuristic L1.     

Comparison of Kernel Discriminant Coordinates Algorithms

The construction of kernel discriminant coordinates reduces to the solution of a generalized eigenvalue problem in which both matrices are nonnegative definite. Six different algorithms for solving that problem are described, and the performance of these algorithms is tested on 26 different datasets. The percentage of misclassifications using a linear discriminant function is noted, and the algorithms’ running times are ascertained. Classification is also performed in the space of classical discriminant coordinates.     

Sample size planning for testing significance of curves

Smoothing methods for curve estimation have received considerable attention in statistics with a wide range of applications. However, to our knowledge, sample size planning for testing significance of curves has not been discussed in the literature. This paper focuses on sample size calculations for nonparametric regression and partially linear models based on local linear estimators. We describe explicit procedures for sample size calculations based on non- and semi-parametric F-tests. Data examples are provided to demonstrate the use of the procedures.     

An adjustable robust optimization model for the resource-constrained project scheduling problem with uncertain activity durations

This paper addresses the resource-constrained project scheduling problem with uncertain activity durations. An adaptive robust optimization model is proposed to derive the resource allocation decisions that minimize the worst-case makespan, under general polyhedral uncertainty sets. The properties of the model are analyzed, assuming that the activity durations are subject to interval uncertainty where the level of robustness is controlled by a protection factor related to the risk aversion of the decision maker. A general decomposition approach is proposed to solve the robust counterpart of the resource-constrained project scheduling problem, further tailored to address the uncertainty set with the protection factor. An extensive computational study is presented on benchmark instances adapted from the PSPLIB.     

Ranked-Weighted Utilities and Qualitative Convolution

For gambles—non-numerical consequences attached to uncertain chance events—analogues are proposed for the sum of independent random variables and their convolution. Joint receipt of gambles is the analogue of the sum of random variables. Because it has no unique expansion as a first-order gamble analogous to convolution, a definition of qualitative convolution is proposed. Assuming ranked, weighted-utility representations (RWU) over gains (and, separately, over losses, but not mixtures of both), conditions are given for the equivalence of joint receipt, qualitative convolution, and a utility expression like expected value. As background, some properties of RWU are developed.     

On the accuracy in high‐dimensional linear models and its application to genomic selection

Genomic selection is today a hot topic in genetics. It consists in predicting breeding values of selection candidates, using the large number of genetic markers now available owing to the recent progress in molecular biology. One of the most popular methods chosen by geneticists is ridge regression. We focus on some predictive aspects of ridge regression and present theoretical results regarding the accuracy criteria, that is, the correlation between predicted value and true value. We show the influence of singular values, the regularization parameter, and the projection of the signal on the space spanned by the rows of the design matrix. Asymptotic results in a high‐dimensional framework are given; in particular, we prove that the convergence to optimal accuracy highly depends on a weighted projection of the signal on each subspace. We discuss on how to improve the prediction. Last, illustrations on simulated and real data are proposed.