基于主成分分析和贝叶斯正则化BP神经网络的GDP预测

选用财政收入、财政支出、消费品零售总额、实际利用外资、进出口总额以及全社会固定资产投资等对GDP有显著影响的6个因子,用1985~2008年中国的宏观经济数据建立了一个基于主成分分析和贝叶斯正则化BP神经网络的预测模型,并把它应用于我国GDP的预测。实证结果表明：通过主成分分析法和贝叶斯正则化方法对BP神经网络进行改进,可简化网络结构,增强泛化能力。与其它常用的预测方法相比,该方法数据输入简便,收敛速度快,拟合曲线光滑,且在预测精度上有明显的优势。     

Effects of deer and mice on seedling survival in a temperate mixed forest of Japan

To detect the magnitude of indirect positive effects of deer and mice on seedling survival of some woody species in a Japanese temperate forest, we analyzed the data from an earlier field experiment using a hierarchical Bayesian approach. The forest studied was inhabited by sika deer (Cervus nippon) and mice (Apodemus spp.), and the floor was covered with dwarf bamboo, Sasa nipponica, which negatively affected tree seedlings and was in turn negatively affected by deer and mice. The field experiment was designed as the combination of exclusion or removal of these factors: deer, mice and dwarf bamboo. A hierarchical Bayesian model was constructed and the parameters were estimated by the Markov chain Monte Carlo method. The model successfully showed the indirect positive effects of deer and mice, which improved the survival rate of five cohorts of the seedlings of three tree species by ameliorating the direct negative effect of dwarf bamboo. The cohorts studied were formed by the seedlings of Abies homolepis that emerged in 1997 and 2002, those of Fraxinus lanuginosa f. serrata that emerged in 1998 and 2002, and those of Fagus crenata that emerged in 1999. The positive indirect effect was especially large in F. crenata, which is known to be severely affected by dwarf bamboo. The total effect of deer was shown to be positive for all cohorts except the A. homolepis cohort that emerged in 1997, at which time the dwarf bamboo had not yet fully recovered from the browsing pressure of deer. The total effect of mice was shown to be positive for all of the cohorts. We conclude that these positive effects were due to the large negative effect of dwarf bamboo on the seedlings.     

Estimation of soil salinity in semi‐arid land using a geostatistical model

Soil salinity is one of the great problems in arid and semi‐arid environments. The estimation and prediction of spatial soil salinity may be considered as a stochastic process, observed at irregular locations in space. Environmental variables usually show spatial dependence among observations which is an important drawback to traditional statistical methods. Geostatistical techniques that analyse and describe the spatial dependence and quantify the scale and intensity of the spatial variation, provides spatial information for local estimation of soil salinity. In this paper we propose a Gaussian Spatial Linear Mixed Model (GSLMM), which involves a non‐parametric term, accounting for a deterministic trend given by exogenous variables, and a parametric component defining the purely spatial random variation, possibly due to latent spatial processes. We focus here on the analysis of the relationship between soil electrical conductivity as a parameter related directly to soil salinity as well as sodium content (sodicity) to identify spatial variations in these parameters. This kind of methodology is demonstrated as a useful tool for environmental land management. Copyright © 2007 John Wiley & Sons, Ltd.     

畜禽基因组选择中贝叶斯方法及其参数优化策略

品种选育在畜禽育种中占十分重要的地位,基因组选择作为畜禽育种的新兴技术手段而备受关注。其优点为可以缩短世代间隔,加快遗传进展,可以不依赖于表型进行选择。2001年,Meuwisen提出基因组选择的概念后,基因组选择首先应用于奶牛育种,至2014年8月,国际公牛组织已有34个成员国在其国家奶牛育种群中应用基因组选择。随着基因组选择的不断推广应用,提高基因组育种值估计准确性的问题有待于解决,当前对基因组选择方法的研究和探讨正在不断深入,有效的模型及算法对提高基因组育种值估计的准确性具有重大现实意义。至今已有17种贝叶斯方法相继被提出,本文简要介绍了基因组选择中的经典BayesA和BayesB方法,其中BayesA假设所有位点都有效应,BayesB假设部分位点有效应,且这部分有效应的位点所占的比例很小,它们的假设模型和算法都不相同。Meuviwisen提出经典贝叶斯方法后,其它贝叶斯方法犹如雨后春笋般涌出,这些新方法的提出,都是基于经典贝叶斯方法原理,对假设模型和算法进行适当改进,以期对模型中的参数进行优化。如BayesC方法在BayesB的基础上对模型中的π值进行优化,BayesCπ和 BayesDπ是在BayesC的基础上进行改进,这两种方法假设各位点的效应方差是相同的,而BayesC假设各位点的效应方差是不同的,BayesDπ又是在BayesCπ基础上对效应方差服从尺度逆卡方分布中的尺度参数进行优化。Bayes Lasso的思想和BayesA一样,不同之处在于它假设标记效应服从另一种分布-拉普拉斯分布,所以标记效应的后验分布也随之改变。BayesRS方法假设各位点的效应方差是按占一定比例的总遗传方差分配的。其它的贝叶斯方法也都是在前人研究的基础之上对模型中的先验假设进行变换和模型中的参数进行优化,以期寻找最适合群体的假设模型和参数。目前广泛应用的贝叶斯算法仍是经典贝叶斯算法以及BayesCπ,这是由于它们计算结果的稳定性和较高的基因组育种值估计准确性。在这3种贝叶斯算法中,基因组育种值估计准确性基本上是BayesB＞BayesCπ＞BayesA,但某些性状计算的基因组育种值准确性结果并非如此。相对于经典贝叶斯方法,参数优化过程在一定程度上提高了基组育种值估计的准确性。总之,在经典贝叶斯方法的基础上,贝叶斯方法的改进算法及其参数优化策略围绕着以提高基因组育种值估计的准确性为目的,通过生物遗传算法与实际的群体情况相结合,寻找最适的假设模型和参数优化策略,丰富和拓展了基因组选择算法,并能使得基因组育种值更具参考价值。由于中国的动物育种历程与国外育种差距甚远,利用基因组选择可以加快畜禽育种进程,进而还可以培养新品系,丰富遗传资源。同时对基因组选择在中国的方法研究及应用进行了介绍,面对基因组选择的种种优点,全基因组选择育种技术势在必行。此外,文章还探讨了畜禽基因组选择中贝叶斯方法及其参数优化策略存在的主要问题和今后研究的热点,以期为获得更加可靠和快捷的基因组选择算法提供参考。     

贝叶斯统计学在遗传中的应用

经典统计分析方法在遗传学中已经得到广泛应用,而在解决许多复杂的遗传学问题中,贝叶斯方法显得更为有效,因而贝叶斯统计学的应用也逐渐受到重视。本文主要讨论贝叶斯统计分析方法与经典统计分析方法的异同、贝叶斯统计学在遗传学的主要应用及其优点。     

Management interventions in dairy herds: Exploring within herd uncertainty using an integrated Bayesian model

Knowledge of the efficacy of an intervention for disease control on an individual farm is essential to make good decisions on preventive healthcare, but the uncertainty in outcome associated with undertaking a specific control strategy has rarely been considered in veterinary medicine. The purpose of this research was to explore the uncertainty in change in disease incidence and financial benefit that could occur on different farms, when two effective farm management interventions are undertaken. Bovine mastitis was used as an example disease and the research was conducted using data from an intervention study as prior information within an integrated Bayesian simulation model. Predictions were made of the reduction in clinical mastitis within 30 days of calving on 52 farms, attributable to the application of two herd interventions previously reported as effective; rotation of dry cow pasture and differential dry cow therapy. Results indicated that there were important degrees of uncertainty in the predicted reduction in clinical mastitis for individual farms when either intervention was undertaken; the magnitude of the 95% credible intervals for reduced clinical mastitis incidence were substantial and of clinical relevance. The large uncertainty associated with the predicted reduction in clinical mastitis attributable to the interventions resulted in important variability in possible financial outcomes for each farm. The uncertainty in outcome associated with farm control measures illustrates the difficulty facing a veterinary clinician when making an on-farm decision and highlights the importance of iterative herd health procedures (continual evaluation, reassessment and adjusted interventions) to optimise health in an individual herd.     

Genetic correlations between categorical morphological traits in Nelore cattle by applying Bayesian analysis under a threshold animal model

In this study, Bayesian analysis under a threshold animal model was used to estimate genetic correlations between morphological traits (body structure, finishing precocity and muscling) in Nelore cattle evaluated at weaning and yearling. Visual scores obtained from 7651 Nelore cattle at weaning and from 4155 animals at yearling, belonging to the Brazilian Nelore Program, were used. Genetic parameters for the morphological traits were estimated by two‐trait Bayesian analysis under a threshold animal model. The genetic correlations between the morphological traits evaluated at two ages of the animal (weaning and yearling) were positive and high for body structure (0.91), finishing precocity (0.96) and muscling (0.94). These results indicate that the traits are mainly determined by the same set of genes of additive action and that direct selection at weaning will also result in genetic progress for the same traits at yearling. Thus, selection of the best genotypes during only one phase of life of the animal is suggested. However, genetic differences between morphological traits were better detected during the growth phase to yearling. Direct selection for body structure, finishing precocity and muscling at only one age, preferentially at yearling, is recommended as genetic differences between traits can be detected at this age.     

Relationships between Podolic cattle breeds assessed by single nucleotide polymorphisms (SNPs) genotyping

Italian Maremmana, Turkish Grey and Hungarian Grey breeds belong to the same Podolic group of cattle, have a similar conformation and recently experienced a similar demographic reduction. The aim of this study was to assess the relationship among the analysed Podolic breeds and to verify whether their genetic state reflects their history. To do so, approximately 100 single nucleotide polymorphisms (SNPs) were genotyped on individuals belonging to these breeds and compared to genotypes of individuals of two Italian beef breeds, Marchigiana and Piemontese, which underwent different selection and migration histories. Population genetic parameters such as allelic frequencies and heterozygosity values were assessed, genetic distances calculated and assignment test performed to evaluate the possibility of recent admixture between the populations. The data show that the physical similarity among the Podolic breeds examined, and particularly between Hungarian Grey and Maremmana cattle that experienced admixture in the recent past, is mainly morphological. The assignment of individuals from genotype data was achieved using Bayesian inference, confirming that the set of chosen SNPs is able to distinguish among the breeds and that the breeds are genetically distinct. Individuals of Turkish Grey breed were clearly assigned to their breed of origin for all clustering alternatives, showing that this breed can be differentiated from the others on the basis of the allelic frequencies. Remarkably, in the Turkish Grey there were differences observed between the population of Enez district, where in situ conservation studies are practised, and that of Bandirma district of Balikesir, where ex situ conservation studies are practised out of the original raising area. In conclusion, this study demonstrates that molecular data could be used to reveal an unbiased view of past events and provide the basis for a rational exploitation of livestock, suggesting appropriate cross-breeding plans based on genetic distance or breeding strategies that include the population structure.     

基于神经网络与贝叶斯信息融合的小白菜成熟度检测方法

为了掌握智能植物生长柜中小白菜的成熟情况,便于对柜内环境参数实现智能控制,提出了利用小白菜的外部形态特征,特别是提取根系形态特征并将其与地上部分形态特征相结合来检测小白菜成熟度的方法。通过Matlab图像处理工具箱对采集的小白菜图像进行阈值分割和特征提取,然后将小白菜上、下两部分的形态特征数据作为训练样本,分别建立对应的神经网络成熟度检测模型,并将神经网络检测值利用贝叶斯理论来对其进行信息融合,从而进一步提高神经网络模型检测的准确性。     

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。