加速寿命数据的贝叶斯建模与分析

在加速寿命试验的可靠性设计中, 随机化设计的限制以及删失数据不可避免地导致低分位数估计出现较大的偏差。针对上述的问题, 结合贝叶斯抽样技术以及非线性混合模型(nonlinear mixed model, NLMM)提出了一种可靠性改进的分析方法。首先, 需要检验所收集的数据是否服从威布尔分布以及验证形状参数是否是恒定常数。其次, 考虑随机效应对尺度参数和形状参数的影响, 运用NLMM构建了尺度参数和形状参数与试验因子之间的函数关系。然后, 利用贝叶斯方法估计低分位数的可靠性寿命。最后, 实际案例研究表明, 在考虑删失问题和未完全随机设计的影响时, 所提方法能够获得更为稳健和可靠的估计结果。     

Description of the aberrant Leptopilina lasallei n. sp., with an updated phylogeny of Leptopilina Förster (Hymenoptera: Figitidae: Eucoilinae)

ABSTRACT

In the search for native Asian parasitoids of Drosophila suzukii, the notorious spotted-wing Drosophila (SWD), an odd new species of Eucoilinae was discovered. Leptopilina lasallei sp. nov. is herein described and diagnosed relative to other eucoilines associated with drosophilid hosts. Morphologically, L. lasallei is somewhat aberrant within Leptopilina; phylogenetically, L. lasallei is sister group to the core Leptopilina. In the process of investigating L. lasallei, a de novo molecular phylogeny of Leptopilina was generated and is included here. The integrated approach used for the characterisation of L. lasallei, and the resulting phylogeny of Leptopilina, produced data useful to select parasitoid species for SWD biological control.http://www.zoobank.org/urn:lsid:zoobank.org:act:402D504A-4616-4524-85D7-1C13A6276F06 http://www.zoobank.org/urn:lsid:zoobank.org:act:402D504A-4616-4524-85D7-1C13A6276F06     

基于时间窗的SVDBN近似推理算法研究

变结构动态贝叶斯网络(SVDBN)在处理非稳态过程的不确定问题具有其独特的优越性。为克服SVDBN推理算法不能实现在线推理的缺陷,在引入SVDBN的时间窗和时间窗宽度概念基础上,阐述了基于时间窗的SVDBN在线近似推理机制,提出了2种在线近似推理算法,即基于时间窗的变结构离散动态贝叶斯网络(SVDDBN)递推推理算法和基于时间窗的SVDDBN快速推理算法。通过仿真实验验证了这2种算法的有效性,并从复杂度、适用范围及更新时间等方面进行了比较。     

Spatial dependence model with feature difference

This paper presents a new spatial dependence model with an adjustment of feature difference. The model accounts for the spatial autocorrelation in both the outcome variables and residuals. The feature difference adjustment in the model helps to emphasize feature changes across neighboring units, while suppressing unobserved covariates that are present in the same neighborhood. The prediction at a given unit incorporates components that depend on the differences between the values of its main features and those of its neighboring units. In contrast to conventional spatial regression models, our model does not require a comprehensive list of global covariates necessary to estimate the outcome variable at the unit, as common macro-level covariates are differenced away in the regression analysis. Using the real estate market data in Hong Kong, we applied Gibbs sampling to determine the posterior distribution of each model parameter. The result of our empirical analysis confirms that the adjustment of feature difference with an inclusion of the spatial error autocorrelation produces better out-of-sample prediction performance than other conventional spatial dependence models. In addition, our empirical analysis can identify components with more significant contributions.     

Dynamic Model Averaging and CPI Inflation Forecasts: A Comparison between the Euro Area and the United States

The paper forecasts consumer price inflation in the euro area (EA) and in the USA between 1980:Q1 and 2012:Q4 based on a large set of predictors, with dynamic model averaging (DMA) and dynamic model selection (DMS). DMA/DMS allows not solely for coefficients to change over time, but also for changes in the entire forecasting model over time. DMA/DMS provides on average the best inflation forecasts with regard to alternative approaches (such as the random walk). DMS outperforms DMA. These results are robust for different sample periods and for various forecast horizons. The paper highlights common features between the USA and the EA. First, two groups of predictors forecast inflation: temporary fundamentals that have a frequent impact on inflation but only for short time periods; and persistent fundamentals whose switches are less frequent over time. Second, the importance of some variables (particularly international food commodity prices, house prices and oil prices) as predictors for consumer price index inflation increases when such variables experience large shocks. The paper also shows that significant differences prevail in the forecasting models between the USA and the EA. Such differences can be explained by the structure of these respective economies. Copyright © 2015 John Wiley & Sons, Ltd.     

Predicting Recessions with Leading Indicators: Model Averaging and Selection over the Business Cycle

Four methods of model selection—equally weighted forecasts, Bayesian model‐averaged forecasts, and two models produced by the machine‐learning algorithm boosting—are applied to the problem of predicting business cycle turning points with a set of common macroeconomic variables. The methods address a fundamental problem faced by forecasters: the most useful model is simple but makes use of all relevant indicators. The results indicate that successful models of recession condition on different economic indicators at different forecast horizons. Predictors that describe real economic activity provide the clearest signal of recession at very short horizons. In contrast, signals from housing and financial markets produce the best forecasts at longer forecast horizons. A real‐time forecast experiment explores the predictability of the 2001 and 2007 recessions. Copyright © 2015 John Wiley & Sons, Ltd.     

基于近似贝叶斯计算方法的排队模型参数估计

基于贝叶斯推断对单服务台的马尔科夫与非马尔科夫系统下的4个排队模型做参数估计与效果评价,并对银行排队叫号换汇数据做实证研究。采用近似贝叶斯计算的方法有效解决复杂排队模型似然函数难解析表达的问题。利用R中queue computer包给出M/M/1,M/G/1,G/M/1,GI/GI/1这4个排队模型的到达时间与服务时间的参数估计及后验分布图。其中M/M/1:估计值(真实值)分别约为1.02(1)、 1.12(1/0.9),都与真实值接近;M/G/1:1.17(1)、1.21(1.2),对服务时间的估计效果优于到达时间;G/M/1:0.49(0.5)、0.96(1),从后验分布进一步得到对服务时间的估计更准确;GI/GI/1:1.16(1.1)、1.07(1)、0.23(0.251)、0.22(0.25),各分量的估计值与真实值相对接近。对于实际银行数据,估计得的参数所拟合的数据与Ausin的拟合数据分布接近。研究表明,近似贝叶斯的方法在排队模型的参数估计上有较大优势,在实际数据中取得了较好应用。     

融合多源先验信息的舰船备件可靠性评估

舰船工作环境常导致备件的实际可靠性规律与设计参数不一致。在实际使用消耗量较小的条件下,全面分析备件可靠性信息来源,通过定义似然权重系数,融合多源先验信息,建立指数型备件可靠性的贝叶斯评估方法。最后,通过仿真分析验证方法的合理性。结果表明,所提出的基于似然权重的先验信息融合方法可以较好地融合不同信息源的可靠性信息,估计方法具有较好的稳定性。     

异常值和未知观测噪声鲁棒的卡尔曼滤波器

给出了对异常值和未知分布的观测噪声鲁棒的卡尔曼滤波器。分析表明当以Huber损失函数替代推导卡尔曼滤波器最大后验准则中观测误差的l₂范数时, 就构造了一个新的准则。由于Huber损失函数可同时描述l₁和l₂范数, 因此由这个新准则推导的卡尔曼滤波器, 在具有传统卡尔曼滤波器性质的同时, 也有了l₁范数对异常值鲁棒的特性。而当含异常值的观测噪声统计分布未知时, 利用含未知参数的高斯混合模型描述其分布以及变分贝叶斯推理, 提出了对异常值和未知统计分布观测噪声鲁棒的卡尔曼滤波器。仿真和实验在验证了分析结果正确的同时, 也表明提出算法的性能优于现有文献报道鲁棒类的卡尔曼滤波器。     

基于变分贝叶斯估计方法的双尺度自适应Kalman滤波

针对Kalman滤波在对敌目标估计应用中遇到的量测和过程噪声均未知且时变的情况,提出了一种利用变分贝叶斯估计的双尺度自适应滤波方法。解决了2个关键问题：一是针对量测和过程噪声协方差的共轭后验分布提出了相对转移概率指标,设计了启发式的自适应噪声估计窗口,实现了稳态精度和时变响应性能的综合提升,能适应敌方目标机动性高且统计特性变化快的特点;二是设计了在不同时间尺度上估计过程噪声和量测噪声的协方差方法,解决了在同一时间尺度上使协方差估计值发生严重偏差且增大滤波误差的问题。仿真表明,所提方法能快速跟踪目标状态噪声统计特性的变化并保证估计精度。