基于压缩主成分估计的两类预测的最优性判别

本文以压缩主成分估计为基础，对广义线性模型的最优预测与经典预测的最优性判别问题进行了讨论，获得了在离差矩阵判别准则和广义风险函数判别准则下判断两类预测量最优性的一个充分条件，为进一步研究基于有偏估计关于两类预测量的最优性判别问题提供了一种方法和思路。     

张量积等距曲面的样条逼近

在几何造型系统中,通常需要用低次有理参数曲线、曲面来逼近等距曲线、曲面.这篇文章主要研究张量积等距曲面的样条逼近.利用样条曲面和原曲面加权组合构造一个新的有理曲面,该曲面通过插值原曲面的等距曲面上的采样点,从而逼近等距曲面.此方法较为简单,逼近曲面的次数不会超过原曲面,逼近曲面能达到C2连续.由插值点决定控制点的个数和逼近所能达到的误差精度,而且可以通过调节权值使等距曲面达到最佳逼近.     

MINIMAX INVARIANT ESTIMATOR OF CONTINUOUS DISTRIBUTION FUNCTION UNDER LINEX LOSS

In this paper we consider the problem of estimation of a continuous distribution function under the LINEX loss function. The best invariant estimator is obtained and proved to be minimax for any sample size n ≥ 1.     

NQD样本下非参数回归函数最近邻密度估计的强相合性

本文在样本为平稳的两两NQD的情况下得到了非参数回归函数m(x)的最近邻估计mn(x)的强相合性.在条件不变的情况下获得了与独立情形一样的mn(x)是强相合的充分条件.     

样条无单元法在弹性薄板振动问题中的应用

本文以结构动力微分方程和三次B样条函数为基础,按一致质量矩阵,推导出样条无单元法分析弹性薄板动力问题的具体计算格式,并编制了相应的计算程序.该方法适用于不同边界条件的弹性薄板的动力特性分析,计算结果表明,本文方法求解薄板动力的特征值问题具有较高精度、易于实施等优点,用较少的结点离散就能获得较好的结果.     

带有不完全椭球约束的多元线性模型中线性估计的可容许性

 对于带有不完全椭球约束的多元线性模型,在参数可估和不可估2种情况下给出了风险函数的6种比较标准,并证明了这些标准是一致的.在齐次线性估计类和非齐次线性估计类中分别给出了参数估计可容许的充要条件.     

广义最小二乘估计的稳健性

讨论了协方差阵未知的椭球等高线性模型中的稳健性问题. 证明当协方差阵在一定范围内变动时, 广义最小二乘估计在一大类损失函数下都是风险最小的估计; 广义最小二乘估计关于协方差阵和损失函数 同时具有稳健性.     

基于双三次样条重建断层扫描图像

断层扫描(CT)技术在医学领域的应用,大大促进了医学的发展。本文采用的双三次样条插值算法是在三次样条插值算法上扩展并成功地应用到CT切片数据当中的。经过插值处理的切片显示给医生,对提高医疗诊断的准确性有很大的帮助。     

Tracking data management using topology of moving object database system for moving object constrained trajectory in LBS application

Advances in wireless technologies and positioning technologies and spread of wireless devices, an interest in LBS (Location Based Service) is arising. To provide location based service, tracking data should have been stored in moving object database management system (called MODBMS) with proper policies and managed efficiently. So the methods which acquire the location information at regular time intervals then, store and manage have been studied. In this paper, we suggest tracking data management techniques using topology that is corresponding to the moving path of moving object. In our techniques, we update the MODBMS when moving object arrived at a street intersection or a curved road which is represented as the node in topology and predict the location at past and future with attribute of topology and linear function. In this technique, location data that are corresponding to the node in topology are stored, thus reduce the number of update and amount of data. Also in case predicting the location,because topology are used as well as existing location information, accuracy for prediction is increased than applying linear function or spline function.     

正交样条小波的周期化

本文对正交样条小波进行周化，给出了周期正交样条小波的分解与重构算法．用这种周期正交样条小波可方便地处理有限区间上的问题．