多项式回归的快速算法

幂函数系在任意有限区间上是封闭的,任一单输入单输出系统的静态特性都能用多项式平均逼近到任意程度.因此,多项式回归在回归分析方法中占有特别重要的地位.做多项式回归时,首先应该确定多项式模型的方次.目前常用的确定方次的方法有F-检验、FPE-检验、AIC-准则、CAT-准则、残差平方和相对变化量检验等.这些检验方法都要用到模型参数最小二乘估计的残差平方和,但是,按照以往的算法,     

小区间内一个素数和三个素数的平方和问题

研究了把一个满足必要条件的自然数在小区间内分解成一个素数和三个素数平方和的问题,利用刘建亚和展涛处理扩大了的主区间的新方法,成功的缩短了小区间的长度.     

非正交试验的平方和分解及其应用

在方差分析中,平方和分解占有很重要的地位.在正交试验里,可以将平方和分解为各效应平方和和误差平方和之和,且使各效应平方和相互独立,从而对各效应假设分别作F 检验.但在实际中,我们经常碰到的数据并不是正交的,此时用最小二乘法得到的估计,其总平方和不再等于各效应平方和与误差平方和之和,而且计算复杂,应用起来极不     

两均匀分布总体参数之比的估计

给出了两均匀分布的最大次序统计量的密度函数,并讨论了两均匀分布参数之比的通常区间估计、最短区间估计及假设检验方法.最后,根据实例求出了这两种区间估计及其区间长度,并得出了相应的结论.     

0-1回归的非参数区间估计

综合非参数方法和Bayes思想提出0-1回归的区间估计方法,并证明了该估计的均方收敛性.通过数值模拟,对各种不同情况下区间估计的覆盖率、相对覆盖率及区间长度进行了对比分析.结果表明:区间估计的效果与样本观测总量有关,样本观测总量越大,覆盖率的波动越稳定,区间长度越短,相对覆盖率也就越大.对于较小的样本量,所得的区间估计仍然具有较好的效果.     

回归系数的岭型主相关估计及其优良性

本文研究了线性模型中的一种有偏估计,利用均方误差和残差平方和,得到了岭型主相关估计的一些性质,是对[1]中相关结果的推广.     

基于β分布的区间估计量化方法

区间估计是专家评价中的重要方法.针对区间估计的特点,构建了区间序列分布,分析了区间序列分布的特性和数字特征;结合β分布及其特性,提出了一种基于β分布的区间估计量化方法,着重探讨了用β分布拟合区间序列分布的原理方法;通过案例分析,表明该方法效果较好,并具有较好的普适性.     

指数分布串—并联混合系统产品的统计分析

给出单元寿命服从同一指数分布的串-并联混合系统产品参数的矩估计和极大似然估计,并通过大量Monte-Carlo模拟比较了估计的精度,得到在样本容量小于35时矩估计优于极大似然估计,而样本容量不小于35时极大似然估计优于矩估计.另外,还给出了参数的精确区间估计与近似区间估计,并通过大量Monte-Carlo模拟考察了区间估计的精度.     

二维均匀分布矩形区域面积的估计

根据一维均匀分布区间长度的区间估计,探讨了二维均匀分布矩形区域面积的估计问题,并给出了其各种估计.     

基于泛岭估计对岭估计过度压缩的改进方法

岭估计是解决多元线性回归多重共线性问题的有效方法,是有偏的压缩估计。与普通最小二乘估计相比,岭估计可以降低参数估计的均方误差,但是却增大残差平方和,拟合效果变差。本文提出一种基于泛岭估计对岭估计过度压缩的改进方法,可以改进岭估计的拟合效果,减小岭估计残差平方和的增加幅度。     

Simultaneous optimality of LSE and ANOVA estimate in general mixed models

Problems of the simultaneous optimal estimates and the optimal tests in general mixed models are considered. A necessary and sufficient condition is presented for the least squares estimate of the fixed effects and the analysis of variance （Hendreson III＇s） estimate of variance components being uniformly minimum variance unbiased estimates simultaneously. This result can be applied to the problems of finding uniformly optimal unbiased tests and uniformly most accurate unbiased confidential interval on parameters of interest, and for finding equivalences of several common estimates of variance components.     

Simultaneous optimal estimates of fixed effects and variance components in the mixed model

For a general linear mixed model with two variance components, a set of simple conditions is obtained, under which, (i) the least squares estimate of the fixed effects and the analysis of variance (ANOVA) estimates of variance components are proved to be uniformly minimum variance unbiased estimates simultaneously; (ii) the exact confidence intervals of the fixed effects and uniformly optimal unbiased tests on variance components are given; (iii) the exact probability expression of ANOVA estimates of variance components taking negative value is obtained.     

On sums of squares of primes II

In this paper we continue our study, begun in G. Harman and A.V. Kumchev (2006) [10], of the exceptional set of integers, not restricted by elementary congruence conditions, which cannot be represented as sums of three or four squares of primes. We correct a serious oversight in our first paper, but make further progress on the exponential sums estimates needed, together with an embellishment of the previous sieve technique employed. This leads to an improvement in our bounds for the maximal size of the exceptional sets.     

Least squares generalized inferences in unbalanced two-component normal mixed linear model

In this paper, we make use of least squares idea to construct new fiducial generalized pivotal quantities of variance components in two-component normal mixed linear model, then obtain generalized confidence intervals for two variance components and the ratio of the two variance components. The simulation results demonstrate that the new method performs very well in terms of both empirical coverage probability and average interval length. The newly proposed method also is illustrated by a real data example.     

On emphasizing variance in the teaching of analysis of variance designs

Many common analysis of variance designs can be rapidly computed with the use of calculators and spreadsheet packages that have standard deviation functions. The calculator methods in this article free students from the tedium of calculating sums of squares and yet enable them to understand how sums of squares are obtained in ANOVA designs.     

因子平方和的正交对照分解及其应用

本文把a水平因子的平方和分解成相互正交的a-1个对照的平方和,这样总变差平方和就可以分解成a个部分(包括残差项),然后又将该分解方法推广到了多因子的情形,并通过因子平方和的分解找到了多因子交互效应对应的对照向量,这使得多水平因子交互效应的计算和解释更加容易,也为方差分析带来了更多的方便,最后给出了几个应用示例。     

具有NA误差项的多元回归模型中最小二乘估计的强相合性

在具有NA误差项的多元回归模型中,对误差项和设计矩阵作了一定的限制条件下给出了其最小二乘估计的强相合性;进而得出了NA序列的样本均值的加权和的几乎处处收敛性的结论.     

Robust Frequency Estimation Using Elemental Sets

Abstract

The extraction of sinusoidal signals from time-series data is a classic problem of ongoing interest in the statistics and signal processing literatures. Obtaining least squares estimates is difficult because the sum of squares has local minima O(1/n) apart in the frequencies. In practice the frequencies are often estimated using ad hoc and inefficient methods. Problems of data quality have received little attention. An elemental set is a subset of the data containing the minimum number of points such that the unknown parameters in the model can be identified. This article shows that, using a variant of the classical method of Prony, parameter estimates for a sum of sinusoids can be obtained algebraically from an elemental set. Elemental set methods are used to construct finite algorithm estimators that approximately minimize the least squares, least trimmed sum of squares, or least median of squares criteria. The elemental set estimators prove able in simulations to resolve the frequencies to the correct local minima of the objective functions. When used as the first stage of an MM estimator, the constructed estimators based on the trimmed sum of squares and least median of squares criteria produce final estimators which have high breakdown properties and which are simultaneously efficient when no outliers are present. The approach can also be applied to sums of exponentials, and sums of damped sinusoids. The article includes simulations with one and two sinusoids and two data examples.     

On the representation of large integers as sums of four almost equal squares of primes

In this paper, we are able to establish two localized results on the conjecture that each large integer congruent to 4 modulo 24 can be written as the sum of four squares of primes. The proof is based on the new estimates for exponential sums over primes in short intervals and a technique to get the asymptotic formula on the enlarged major arcs in the circle method. This work is supported by the National Natural Science Foundation of China (Grant Nos. 10701048 and 10771127).