结合多尺度空间滤波和层级网络的基于结构保持的高光谱特征选择

为了充分利用高光谱图像蕴含的丰富的光谱信息和空间信息,提出了结合多尺度空间滤波和层级网络的基于结构保持的高光谱特征选择算法.算法利用基于l2,1范数的数学模型,选出同时保存全局相似性结构和局部流形结构的特征子集;在多个尺度的窗口中使用双边滤波,自适应计算滤波核,自动在光谱数据中融入空间信息,增强了类内相似性和类间相异性,避免了参量选择;引入层级结构实现空间信息和光谱信息的深入融合,提高了分类准确度;讨论了层级数目和窗口尺度个数对分类准确度的影响.在Indian Pines和PaviaU两个数据集的实验表明,该算法在大部分地物种类上的分类准确度都有较大幅度的提升,总体分类准确度分别达到90.98%和94.20%,相比其他方法明显提高了地物分类准确度.     

On the critical nonlinear damped wave equation with large initial data

We study the nonlinear damped wave equation

(0.1)     

��Ⱦ��������ض�������ģ�͵IJ�������

本文研究了线性模型响应变量被污染且被区间截断下的参数估计问题, 借助于区间数据的无偏转换, 得到了回归系数和污染系数的估计, 并在一定的条件下得到了这些估计的强相合性. 通过若干模拟例子说明, 尽管数据经过污染和区间截断的双重信息损失, 但用本文提出的方法得到的估计, 仍能取得良好的估计效果     

Micellar Liquid Chromatography Study of Quantitative Retention–Activity Relationships for Antihypertensive Drugs

Use of micellar mobile phases in reversed-phase liquid chromatography (RPLC) results in hydrophobic and electrostatic sites for interaction. Modified stationary phases in micellar liquid chromatography (MLC) are structurally similar to biomembranes. To confirm this we focused on the effects of the type and concentration of surfactant (Brij 35, SDS, and CTAB) and mobile phase pH on the retention of antihypertensive drugs on modified C₁₈ stationary phases. Quantitative retention-activity relationships are proposed for the drugs and the different surfactants and compared with those obtained using aqueous–organic mobile phases. Finally, a correlation was obtained between the logarithm of retention factors (log k) and the toxicity (LD₅₀) of antihypertensive drugs. Revised: 14 September 2005 and 4 April 2006     

质子与裂变核238U反应的理论计算

利用球型光学模型、耦合道理论、核内级联、激子模型、蒸发模型、带宽度涨落修正的Hauser-Feshbach模型,Bohr-Wheeler裂变理论,对入射能量从阈能到200 MeV的p+²³⁸U各种反应进行了计算,计算结果和实验符合较好.     

Buckley–James-type of estimators under the classical case cohort design

We consider the estimation problem with classical case-cohort data. The case-cohort design was first proposed by Prentice (Biometrics 73:1–11, 1986). Most studies focus on the Cox regression model. In this paper, we consider the linear regression model. We propose an estimator which extends the Buckley–James estimator to the classical case-cohort design. In order to derive the BJE, there is an additional problem of finding the generalized maximum likelihood estimator (GMLE) of the underlying distribution functions. We propose a self-consistent algorithm for the GMLE. We also justify that the GMLE is consistent and asymptotically normally distributed under certain regularity conditions. We further present some simulation results on the asymptotic properties of the BJE and apply our procedure to a data set used in the literature.     

Critical exponent for semilinear damped wave equations in the N-dimensional half space

We generalize a previous result of Ikehata (Math. Methods Appl. Sci., in press), which studies the critical exponent problem of a semilinear damped wave equation in the one-dimensional half space, to the general N-dimensional half space case. That is to say, one can show the small data global existence of solutions of a mixed problem for the equation u_tt−Δu+u_t=|u|^p with the power p satisfying p∗(N)=1+2/(N+1)<p?N/[N−2]+ if we deal with the problem in the N-dimensional half space.     

Validating DEA as a ranking tool: An application of DEA to assess performance in higher education

There is a general interest in ranking schemes applied to complex entities described by multiple attributes. Published rankings for universities are in great demand but are also highly controversial. We compare two classification and ranking schemes involving universities; one from a published report, ‘Top American Research Universities’ by the University of Florida's TheCenter and the other using DEA. Both approaches use the same data and model. We compare the two methods and discover important equivalences. We conclude that the critical aspect in classification and ranking is the model. This suggests that DEA is a suitable tool for these types of studies.     

A class of weighted estimators for additive hazards model in case-cohort studies

Case-cohort sampling is a commonly used and efficient method for studying large cohorts. In many situations, some covariates are easily measured on all cohort subjects, and surrogate measurements of the expensive covariates also may be observed. In this paper, to make full use of the covariate data collected outside the case-cohort sample, we propose＇a class of weighted estimators with general time-varying weights for the additive hazards model, and the estimators are shown to be consistent and asymptotically normal. We also identify the estimator within this class that maximizes efficiency, and simulation studies show that the efficiency gains of the proposed estimator over the existing ones can be substantial in practical situations. A real example is provided.     

The impact of special days in call arrivals forecasting: A neural network approach to modelling special days

A key challenge for call centres remains the forecasting of high frequency call arrivals collected in hourly or shorter time buckets. In addition to the complex intraday, intraweek and intrayear seasonal cycles, call arrival data typically contain a large number of anomalous days, driven by the occurrence of holidays, special events, promotional activities and system failures. This study evaluates the use of a variety of univariate time series forecasting methods for forecasting intraday call arrivals in the presence of such outliers. Apart from established, statistical methods, we consider artificial neural networks (ANNs). Based on the modelling flexibility of the latter, we introduce and evaluate different methods to encode the outlying periods. Using intraday arrival series from a call centre operated by one of Europe’s leading entertainment companies, we provide new insights on the impact of outliers on the performance of established forecasting methods. Results show that ANNs forecast call centre data accurately, and are capable of modelling complex outliers using relatively simple outlier modelling approaches. We argue that the relative complexity of ANNs over standard statistical models is offset by the simplicity of coding multiple and unknown effects during outlying periods.