纵向资料回归系数的新估计

本文研究纵向资料中回归系数的估计,借助于相依回归模型中的信息迭加法,给出了一个新估计,它改进了最小二乘估计．     

相依模型参数估计的相对效率

本文考虑相依模型,对其未知参数向量给出了其最佳线性无偏估计相对于协方差改进估计的四种相对效率,同时,还给出了最小二乘估计相对于协方差改进估计的三种相对效率,在不同条件下,分别给出了相对效率的上界与下界。     

基于总体最小二乘方法的基因表达缺失数据估计

在基因芯片实验中,数据缺失客观存在,并在一定程度上影响芯片数据后续分析结果的准确性。在不增加实验次数的情况下,缺失值估计是降低缺失数据对后续分析影响的有效方法。针对基因表达数据含有噪声的特点,提出了基于总体最小二乘估计的基因表达缺失值估计算法。实验结果表明,新的估计算法具有比传统缺失值估计算法更好的稳定性和估计准确度。     

二次等距抽样

本文提出了一种二次等距抽样方法,并提出了总体平均值的一个估计量：拼配部分的比型估计与轮换部分的样本均值的加权平均。当样本量较大时,求出了估计量的方差及最优轮换比．并对特殊情形进行了讨论和数值比较．     

生物量精确估算模型与参数辨识方法及应用

从生物量模型的构建与参数辨识方法的改进对生物量进行精确估算。用Chebyshev多项式系的组合构建了p维连续函数空间的一组乘积型基,进而建立了生物量估算统一模型,它具有如下特点:(1)可以克服常用生物量估算模型的经验性、不稳定性、不通用性及对生物量影响因素适应性差的特点,(2)它适合于影响生物量的任何因素,故适应范围非常广且很稳定,(3)可根据实际需要及估算精度确定影响生物量的因素及其阶数大小,(4)模型对生物量的估算相当于在区间[-1,1]上进行的数值插值,变量阶数越高,所插入的点就越多,估算结果越符合实际,整个估算的插值过程与树木的树干解析与树木生长原理是相一致的。对所建模型的参数辨识方法做了探讨,经典最小二乘算法是生物量估算的最常用参数辨识方法,由于它本身固有的一些缺陷使常用最小二乘的估算精度与使用范围受到很大的限制,现代多元统计分析的偏最小二乘算法可以克服常用最小二乘的缺陷,但在提取成分时仍具有不足,针对偏最小二乘的缺陷本文对它做了改进,改进算法即能克服偏最小二乘的不足还能使估算精度大大提高。用2个案例对3种生物量估算方法做了对比分析,结果表明生物量估算统一模型与偏最小二乘改进算法精度最高,其生物量估计误差在零附近排成一条直线。     

推广的生长曲线模型的协方差不变最小二乘无偏估计(Ⅰ)

本文用投影理论及矩阵的谱分解方法,找出了推广的生长曲线模型的未知协方差矩阵∑及tr(C∑)在一定条件下的不变量小二乘无偏估计,为以后更深入的研究奠定基础。     

三种回归分析方法在Hyperion影像LAI反演中的比较

借助GPS进行地面精确定位,利用LAI-2000冠层分析仅在攸县黄丰桥林场开展130个样地(60m×60m)的叶面积指数(Leaf Area Index,LAI)测量.采用FLAASH模块对Hyperion数据进行大气校正并与地面同步冠层观测数据进行拟合,通过研究地面实测LAI与Hyperion影像波段及其衍生的系列植被指数(NDVI、RVI等)的相关性,筛选出估算叶面积指数的植被指数因子.应用曲线估计、逐步回归及偏最小二乘三种回归分析技术分别建立叶面积指数的最优估算模型.结果表明:参与建模的因子中,比值植被指数(RVI)与LAI的相关性最大,敏感性最高,其次是SARVI0.1,NDVI705,NDVI,SARVI0.1,SARVI0.25;曲线估计、逐步回归分析和偏最小二乘回归三种分析方法所建的6个回归模型中,偏最小二乘回归的拟合效果最好,预测值与实测值的决定系数R2为0.84、曲线估计的拟合效果最低,预测值与实测值的决定系数R2为0.64;建模精度分析表明,选用5-6个自变量因子进行LAI建模是可靠的,以6个植被因子建立的偏最小二乘回归模型预测精度最高.     

利用多重抽样估计稀有物种的种类数

利用矩估计和二个稳健估计方法（jackknife估计,bootstrap估计）来处理野外生态学工作者的调查数据,在假定已经发现一些稀有物种的情形下,通过统计推断得到那些未被发现的物种的种类数。利用本文所提出的方法调查水稻水稻田的昆虫群落和林地的在面植被群落的稀有种是十分有效的。     

经纬杂交试验分析新法—一般最小二乘

经纬杂交试验主要涉及地方品种。这些品种具有两个显著特点:一是品种之内纯合程度较差,二是品种之间纯合程度不同。所以对于不同杂交组合而言,相应于观测值的误差效应不仅很大而且服从不同的分布;此时若再利用正规最小二乘进行分析,必然不能得到正确结果。本文根据客观特点,依据新的模型和最优线性无偏估计原理,对于这种试验提出了一种新的分析方法——一般最小二乘分析。这种方法只需首先估计各个组合的方差,然后用它对重复数和观测值进行校正,再按正规最小二乘分析步骤分析即可,所以并不复杂。具体使用有一实例说明。     

基于最小二乘支持向量机的农田水汽通量建模

支持向量机是在统计学习理论基础上发展起来的一种新型的机器学习方法,已在模式识别、非线性建模等领域中得到了应用．本文将最小二乘支持向量机方法应用于农田水汽通量的建模中,并同前馈反向传播神经网络的建模性能进行了比较．结果表明,最小二乘支持向量机方法具有可调参数少、学习速度较快等优点,具有更好的推广能力,以更高的精度建立农田水汽通量模型．模型的敏感性分析进一步显示,用最小二乘支持向量机方法建立的农田水汽通量模型是合理可行的．     

Recent advances in the study of Kaposi's sarcoma-associated herpesvirus replication and pathogenesis

It has now been over twenty years since a novel herpesviral genome was identified in Kaposi's sarcoma biopsies. Since then, the cumulative research effort by molecular biologists, virologists, clinicians, and epidemiologists alike has led to the extensive characterization of this tumor virus, Kaposi's sarcoma-associated herpesvirus(KSHV; also known as human herpesvirus 8(HHV-8)), and its associated diseases. Here we review the current knowledge of KSHV biology and pathogenesis, with a particular emphasis on new and exciting advances in the field of epigenetics. We also discuss the development and practicality of various cell culture and animal model systems to study KSHV replication and pathogenesis.     

The structure, reproduction and taxonomy of Vidalia and Osmundaria (Rhodophyta, Rhodomelaceae)

Comprises species occurring mostly in subtidal habitats in tropical, subtropical and warm-temperate areas of the world. An analysis of the type species, V. spiralis (Sonder) Lamouroux ex J. Agardh, a species from Australia, establishes basic characters for distinguishing species in the genus. These characters are (1) branching patterns of thalli, (2) flat blades that may be spiralled on their axis, (3) width of the blade, (4) primary or secondary derivation of sterile and fertile branchlets and (5) position of sterile and fertile branchlets on the thalli. Application of the latter two characters provides an important basic method for separation of species into three major groups. Osmundaria , a genus known only in southern Australia, was studied in relation to Vidalia , and its separation from the Vidalia assemblage is not accepted. Species of Vidalia therefore are transferred to the older genus name, Osmundaria. Two new species, Osmundaria papenfussii and Osmundaria oliveae are described from Natal. Confusion in the usage of the epithet, Vidalia fimbriala Brown ex Turner has been clarified, and Vidalia gregaria Falkenberg, described as an epiphyte on Osmundaria pro/ifera Lamouroux, is revealed to be young branches of the host, Osmundaria prolifera.     

New or rare chromosome counts in Artemisia L. (Asteraceae, Anthemideae) and related genera from Kazakhstan

Fifteen chromosome counts of six Artemisia taxa and one species of each of the genera Brachanthemum, Hippolytia, Kaschgaria, Lepidolopsis and Turaniphytum are reported from Kazakhstan. Three of them are new reports, two are not consistent with previous counts and the remainder are confirmations of very scarce (one to four) earlier records. All the populations studied have the same basic chromosome number, x = 9, with ploidy levels ranging from 2x to 6x. Some correlations between ploidy level, morphological characters and distribution are noted.     

肝癌中HBV和HCV基因和抗原的分布及意义

采用原位分子杂交方法检测HCV RNA及HBV X基因;采用免疫组织化学方法研究HCV核心抗原,非结构区C33c抗原及HBxAg在肝细胞肝癌中的定位及分布.结果表明(1)HCV RNA、HBV X基因在肝细胞肝癌组织检出率分别为40%(55/136)和82%(112/136).HCV RNA定位于癌细胞的胞浆内,阳性细胞呈散在、灶状及弥漫分布三种形式;HBV X基因在肝癌细胞中的分布呈胞浆型、核型及核浆型,阳性细胞也呈上述三种分布形式;(2)HCV C33c抗原、核心抗原在肝细胞肝癌中的阳性率为81%(133/164)及86%(141/164).C33c抗原定位于癌细胞及肝细胞的胞浆内;核心抗原既定位于癌细胞核中,又可定位于胞浆中.C33c抗原阳性细胞以灶状分布为主;而核心抗原阳性细     

Selection with two alleles and complete dominance

For a plant selection model with frequency-independent viabilities, fertilities and selfing rates, it is shown that apart from global fixation, for certain parameter combinations a protected polymorphism and facultative fixation (either allele may become fixed according to initial frequencies) may both occur. Facultative fixation requires different selling rates for the dominant and recessive type. Protection of the polymorphism requires resource allocation for male and female function. In this connection the problem of purely genetically caused population extinction is discussed.
For general frequency dependence and regular segregation, the chances for establishment of a completely recessive gene are compared to those of a completely dominant gene. It is proven that the process of establishment of the recessive gene, despite a fitness advantage, may be considerably endangered by drift effects if random mating prevails. The recessive gene may reach the same effectivity in establishment as a dominant gene, only if the recessive homozygote mates exclusively with its own type during the period of establishment.