ESTIMATION FOR NONNEGATIVE AUTOREGRESSIVE PROCESSES WITH AN UNKNOWN LOCATION PARAMETER

Abstract. Consider an AR(1) process given by X ^t=γ+ø X _t+ Z _t≥ 1. where 0 ≤γ, 0 ≤ø 1 are unknown parameters and the innovations Z _t, ≥ 1, are independently and identically distributed positive random variables. We propose estimates of (γø) which are obtained as the solution to a linear programming problem and establish their strong consistency. When the Z _ts have the exponential distribution. our estimate becomes the conditional maximum likelihood estimate given X ₀. Under the assumption of regular variation of the innovation distribution at its left and right endpoints (assumed to be 0 and ∝ respectively), we establish asymptotic limit laws for the estimates. Consistent estimators for a class of moving-average processes with heavy-tailed innovation distribution are also presented.     

An EM-based identification algorithm for a class of hybrid systems with application to power electronics

In this paper we present an identification algorithm for a class of continuous-time hybrid systems. In such systems, both continuous-time and discrete-time dynamics are involved. We apply the expectation-maximisation algorithm to obtain the maximum likelihood estimate of the parameters of a discrete-time model expressed in incremental form. The main advantage of this approach is that the continuous-time parameters can directly be recovered. The technique is particularly well suited to fast-sampling rates. As an application, we focus on a standard identification problem in power electronics. In this field, our proposed algorithm is of importance since accurate modelling of power converters is required in high- performance applications and for fault diagnosis. As an illustrative example, and to verify the performance of our proposed algorithm, we apply our results to a flying capacitor multicell converter.     

THIRD-ORDER ASYMPTOTIC PROPERTIES OF ESTIMATORS IN GAUSSIAN ARMA PROCESSES WITH UNKNOWN MEAN

Abstract. This paper deals with the third-order asymptotic theory for Gaussian autoregressive moving-average (ARMA) processes with unknown mean μ. We are interested in the estimation of ρ = ( α₁…, α_p, β₁…, β_q ), where α ₁…, α_ρ and β ₁…, β_q are the coefficients of the autoregressive part and the moving-average part, respectively. First, we investigate the third-order asymptotic optimality of the bias adjusted maximum likelihood estimator (MLE) of ρ in the presence of the nuisance parameters μ and ² (innovation variance). Next, for a Gaussian AR(1μ μ, ²), we propose a mean corrected estimator α_c1c2 of the autoregressive coefficient. We make a comparison between the bias adjusted estimator α_c1c2* and the bias adjusted MLE, in terms of their probabilities of concentration around the true value, or equivalently, in terms of their mean squared errors. Finally some numerical studies are provided in order to verify the third-order asymptotic theory.     

最小相关性最大依赖度属性约简

在经典粗糙集中,基于重要度的决策表属性约简算法只考虑了决策属性与条件属性之间的依赖度,没有考虑约简中条件属性之间的相关性,由此求出的约简中可能依然包含冗余属性。针对这一问题,提出了一种改进算法,它利用最小相关性和最大依赖度准则求决策表属性约简。与基于重要度的决策表属性约简算法相比,本算法求出的约简包含的属性个数少、冗余小。实验结果显示,本算法优于基于重要度的决策表属性约简算法。     

一种新的基于最大边缘准则的监督流形学习方法

在深入研究局部样条嵌入算法(LSE)的基础上,引入明确的线性映射关系,构建平移缩放模型和正交化特征子空间,提出了一种正交局部样条判别投影算法(O-LSDP),有效解决了原始LSE算法存在的两个主要问题:样本外点学习问题和无监督模式学习问题。该算法能够应用于模式分类问题并显著改善算法的分类识别能力。在标准人脸数据库上进行的实验比较分析验证了该算法的有效性与可行性。     

准完全最大距离伪随机测试研究

对超大规模集成电路进行随机测试的测试码之间的距离作了定量分析,在此基础上,改进了最大距离随机测试算法中测试码的生成方法,使得所生成的伪随机测试码集合同时达到最大海明距离与近似最大笛卡尔距离。因此每一个测试码可以尽可能多地独立检测到更多不重复的故障。进而提出了准完全最大距离测试新算法的思想和构建理论,并详细阐述了该算法的执行流程。在ISCAS'85基准电路上进行的大量实验数据分析表明,本方法确实有效地提高了随机测试效率,降低了随机测试成本。     

一种融合视觉决策理论预测用户选择的方法

为有效解决在眼动人机交互环境中由于难以判断用户决策而导致的米达斯接触问题, 结合视觉决策理论, 通过融合眼动指标与时间序列来预测用户行为。先记录用户浏览目标页面时的眼动数据及进行选择时的点击数据, 并对数据按时间轴进行采样, 得到注视序列; 然后根据注视序列拟合用户进行决策时的注视似然曲线, 确定相关参数与触发阈值。眼动实验结果表明, 单凭注视阈值预测准确率就可以达到78. 3%, 在人机交互环境中拟合的似然曲线与视觉决策理论耦合。方法的实施无须用户进行额外的配合, 不依赖预设的固定阈值; 相对于设置固定触发阈值, 或要用户进行眨眼、按键等操作的传统方法, 交互过程可更顺畅、自然。     

EFFICIENT DESIGNS FOR NONLINEAR MODELING

In the modeling of physical systems, such as chemical reactions, nonlinear models in a set of unknown parameters are commonly developed. Using the method of least squares for model fitting the efficiency of the process is measured by the large sample variance of the parameter estimates. Designs which yield desirable properties in the linear case, such as factorial and central composite designs, are often utilized in the nonlinear model setting. In this paper a method for transforming these standard designs, yielding more efficient estimators, in the nonlinear model case is proposed. The procedure is demonstrated on a typical kinetic reaction model.     

具有高粘度分散相的油-水分散体系的液滴破碎与粒径分布研究

对湍流搅拌槽中原油-水分散体系的液滴破碎进行了研究。在不同的温度下原油表现出不同的流变学特征，对分散过程中的液滴破碎产生不同的影响。实验研究对比了不同温度下原油-水分散体系的液滴分布及最大稳定粒径，分析了触变性对破碎过程及最大稳定粒径的影响。经过模型计算与实验结果比较，发现以初始粘度计算的理论值预测最大稳定粒径更为合适。     

On Hypotheses Testing for the Selection of Spatio‐Temporal Models

Abstract. Several models have been proposed in recent years for analysing spatial data and also, to some extent, spatio‐temporal data. One of the important problems, namely the choice of an appropriate model for describing real data sets, remains unsolved. Here we consider the analysis of spatio‐temporal processes from which observations over space and time are available. We propose statistical tests for discriminating between space–time autoregressive processes and multivariate autoregressive processes. The sampling properties of the proposed tests are considered. We illustrate the methods with a real example. We use the above tests to find the best model to describe spatio‐temporal variations of hourly carbon monoxide measurements at four locations in London in January 2004.