使用惩罚距离进行说话人改变检测

在说话人改变判决中,为了避免使用贝叶斯信息准则（Bayesian information criterion,BIC）需要不断调节惩罚因子的问题,作者从BIC中推导出惩罚距离公式并将其应用于说话人改变检测中．试验结果表明,提出的算法和BIC相比不需要调节参数,与BIC距离相比,F1提高了2％．     

系统发生分析程序MrBayes 3.1使用方法介绍

在介绍MrBayes3.1程序基本特点以及Nexus文件准备的基础上,选取普通DNA序列、普通蛋白质序列、含编码区域的DNA序列、mRNA序列以及混合型数据文件为例分别介绍了MrBayes3.1程序的基本使用方法,为初学者正确使用该程序提供了操作指南,同时为深入学习与掌握该程序的特殊用途打好基础。     

Bayesian classification of ripening stages of tomato fruit using acoustic impact and colorimeter sensor data

In this work, the concept of data fusion is applied to nondestructive testing data for classification of fresh intact tomatoes based on their ripening stages. A Bayesian classifier considering a multivariate, three-class problem was incorporated for data fusion. Probability of error was estimated numerically for univariate and multivariate cases based on Bhattacharyya distance. Numerical results showed that multi-sensorial data fusion reduces the classification error considerably. The Bayesian classifier was tested on data of tomato fruits taken by the following nondestructive tests: colorimeter and acoustic impact. Results of Bayesian classifier agree with numerical estimations showing an 11% classification error in the multivariate (multi-sensor) case compared with a 48% obtained by the univariate case (single sensor).     

基于贝叶斯与元学习的氨氮浓度预测模型优化

针对小样本数据下氨氮浓度模型预测精度不高、收敛速度较慢的问题,采用长短期记忆网络（long short-term memory,LSTM）建立氨氮浓度预测模型,并利用贝叶斯优化算法和元学习机制对模型进行优化。其中贝叶斯优化算法用来优化预测模型的超参数,同时给出模型参数的初始值,再使用Meta-LSTM算法学习模型梯度并允许优化器之间进行参数共享和更新,最终实现对氨氮浓度预测模型的优化。将该方法与LSTM、GRU和RNN模型进行对比试验,结果显示,研究所建模型对氨氮浓度预测的均方根误差、平均绝对误差和均方误差分别为0.027 6、0.023 9和0.000 76,均优于其他预测模型。表明基于贝叶斯和元学习的氨氮浓度预测模型对小样本数据建模有效,可以实现网络快速收敛,精度满足水产养殖中氨氮浓度预测需求。     

基于改进SHIPP模型的大型原油库区事故风险预测

隐患、未遂事故及无伤亡事故等异常事件是重大事故的早期预警和征兆,此类事件发生频率高,通过建立事故模型识别及纠正事件中的不安全因素能够有效预防重大事故发生。结合油库工艺特点和事故特征,对系统危害辨识、预测及预防(System Hazard Identification,Prediction and Prevention,SHIPP)模型改进,建立基于安全屏障的油库事故模型。采用故障树和事件树相结合的方式构建原因-后果关系模型,将故障树和事件树映射为贝叶斯网络以表征不确定性和条件依赖性。针对新的证据信息,通过贝叶斯网络更新机制实施概率更新;基于贝叶斯理论对现场异常事件数据进行经验学习,降低先验概率的不确定性,实现对油库事故的动态风险预测。对某油库算例分析结果表明,库区发生物质和能量释放的概率较大,整体安全性能趋于退化,应加强安全检查和隐患排查的力度。研究成果可为大型油库风险预测和事故预防提供科学指导和决策支持。(图12,表6,参37)     

荷耦合作用下沥青混凝土疲劳寿命

针对我国通货膨胀水平与不确定性的时变性特征,分别建立了随机波动均值模型和非对称随机波动均值模型,在MCMC稳态模拟的框架下研究了我国通货膨胀水平与不确定性的动态关系。研究结果表明：我国通货膨胀不确定性中具有明显的持续性特征,并且通胀水平中虽然不存在与金融资产价格运动类似的杠杆效应,但是正向冲击增加了经济行为主体对未来不确定性的预期,由此将导致明显的“示范效应”和“追涨效应”;特别是风险溢出系数的贝叶斯估计为正,反映了通胀不确定性对通胀水平的正向影响作用,说明我国目前的货币政策框架中含有相机抉择的成分因素。     

Combination forecasting method of BTP in blast furnace under the conditions of small samples and poor information

The iron ore sintering process is a complex object with the characteristics of uncertainty, multivariable coupling, time-varying and time-delay. The burning-through-point (BTP）is affected by many factors and difficult to be controlled to the required precision by conventional control methods. A BTP prediction method is proposed by using the excellent time series prediction performance of support vector machines (SVMs), and the characteristic that Bayesian theory can make use of sample information and prior knowledge to simplify prediction model and optimize parameters．Firstly, the mechanism of BTP is analyzed, the Bayesian theory and LS-SVM are elaborated respectively, and the Bayesian evidence framework is applied to least squared support vector machine(LS-SVM） regression in order to infer non-linear models for predicting a time series．On the first level of inference, model parameters are selected and on the second level the hyper-parameters are selected．The kernel parameter are tuned on the third level framework,and on this level the relevant inputs are selected．A LS-SVM model is proposed on the basis of the Bayesian LS-SVM models. The results reveal that the BTP of sinter can be accurately predicted by this model even with small samples and poor information. It is concluded that the LS-SVM model is effective with the advantages of high precision, less samples required and simple calculation.     

Estimation of (co)variance components for birth weight in Nellore cattle using alternative data editing criteria

The aim of this study was to estimate the (co)variance components and breeding values for birthweight (BW) in Nellore cattle by considering or not identical weights that exhibit a high frequency within the contemporary group (CG). A total of 175,258 BW records of Nellore cattle born between 2002 and 2018 were used. The CG was formed by farm, year of birth, sex and feeding regime at birth. CGs with more than 16% of identical BW values were eliminated, generating a data file called BWd. Another file was created without removing these animals (BWt). A mixed linear model was used for statistical analysis, which included fixed and random effects. In both data files analysed, single-trait analysis was performed by Bayesian inference. The mean direct and maternal heritability for BW and the correlation between direct and maternal effects were 0.27, 0.07 and −0.07 for BWt, respectively, and 0.30, 0.093 and −0.07 for BWd. This method should affect the estimation of genetic merits of animals for BW, providing greater safety in the choice of sires.     

Modelling nesting habitat as a conservation tool for the Eurasian black vulture (Aegypius monachus) in Dadia Nature Reserve, northeastern Greece

The Eurasian black vulture (Aegypius monachus) is a globally endangered species, vulnerable in Europe and endangered in Greece. We modelled its nesting preferences in Dadia reserve, northeastern Greece using logistic models at multi-scale level combined by Bayesian statistics. Compared with the random sites, the vultures nested at trees with greater diameter (DBH), lower height, and lower total number of trees around the nest tree, steeper slopes and greater distance from forest roads. Our results indicate that conservation and management guidelines for black vultures must aim: (1) to preserve old, isolated mature trees in the reserve, (2) to maintain the zones of strict protection as vulture sanctuary, including most of the suitable nesting habitat, (3) to protect the apparently suitable but still unused habitat in the rest of the reserve from disturbance and logging, (4) to monitor any changes in habitats and use our models to evaluate and predict their effect on vulture nesting in the reserve.     

Parameter estimation in selected populations with missing data

This study proposes a procedure to estimate genetic parameters in populations where a selection process results in the loss of an unknown number of observations. The method was developed under the Bayesian inference scope following the missing data theory approach. Its implementation requires slight modifications to the Gibbs sampler algorithm. In order to show the efficiency of this option, a simulation study was conducted.