回归分析系统的可靠性理论

根据回归分析基本理论及测量平差理论，讨论了观测误差与观测值残差的关系，指明了残差协因素矩阵及多余观测分量在可靠性理论研究中的特殊作用。并在单个一维备选假设下，将可靠性理论的研究方法引入回归分析系统。提出了回归计算系统的内部可靠性及外部可靠性的概念，以及衡量它们的量化尺度。为在大坝资料中分析中剔除粗差，合理筛选观测数据，提供了理论依据。析     

Neural network models for breast cancer prognosis

Estimating the risk of relapse for breast cancer patients is necessary, since it affects the choice of treatment. This problem involves analysing data of times to relapse of patients and relating them to prognostic variables. Some of the times to relapse will usually be censored.We investigate various ways of using neural network models to extend traditional statistical models in this situation. Such models are better able to model both non-linear effects of prognostic factors and interactions between them, than linear logistic or Cox regression models. With the dataset used in our study, however, the prediction of the risk of relapse is not significantly improved when using a neural network model. Predicting the risk that a patient will relapse within three years, say, is possible from this data, but not when any relapse will happen.     

A pretest for using logrank or Wilcoxon in the two-sample problem

In a two-sample location-scale model with censored data, the logrank test is asymptotically efficient when the error distribution is extreme minimum value. On the other hand, the Wilcoxon test is asymptotically efficient when the error distribution is logistic. We propose a pretest for choosing between logrank and Wilcoxon by determining if the error distribution is closer to extreme minimum value or logistic. This adaptive test is compared with the logrank and Wilcoxon tests through simulation.     

A unified view on lifetime distributions arising from selection mechanisms

In this paper, we formulate a flexible density function from the selection mechanism viewpoint (see, for example, Bayarri and DeGroot (1992) and Arellano-Valle et al. (2006)) which possesses nice biological and physical interpretations. The new density function contains as special cases many models that have been proposed recently in the literature. In constructing this model, we assume that the number of competing causes of the event of interest has a general discrete distribution characterized by its probability generating function. This function has an important role in the selection procedure as well as in computing the conditional personal cure rate. Finally, we illustrate how various models can be deduced as special cases of the proposed model.     

Information of group-correlations in Korean financial market

We study two sides of the KOSPI, classified as an emerging market. First, the evolutionary property is examined in terms of overlapping matrix and survival ratios. To this end, we apply the random matrix theory (RMT) and the one-factor model to analyzing correlation matrix and finding business clusters. Second, we examine the relations between the market capitalization and the business. For the well-developed markets such as NYSE, the contribution of the firms to the second-largest eigenvector shows an exponential function of the market capitalizations while no clue is observed for the KOSPI. We confirm that the market capitalization is distributed in a power-law with the exponent 1.2 like a Pareto's distribution. Particulary, the KOSPI shows a different behavior compared to the mature market, that is, one or two companies lead a number of companies with the little money and big companies competed to win each other. The clusters also represent by largest eigenstates show a weak affiliation compared to smaller ones. These results imply that the KOSPI is the target for the short-positioned investors.     

Additive two-way hazards model with varying coefficients

The paper considers smooth modelling of hazard functions, where dynamics is modelled in both, duration time and calendar time. The model is specified with time dynamic covariate effects to replace restrictive assumptions of proportional hazards. Additivity of the time effects is assumed which allows for simple estimation in a backfitting style. Penalized splines are employed, which provide the welcome benefit of linking smoothing with mixed models. The model is applied to unemployment data taken from the German socioeconomic panel. The hazard function, here the chance for finding reemployment, varies with duration as well as calendar time.     

Estimation from aggregate data

A statistical methodology to handle aggregate data is proposed. Aggregate data arise in many fields such as medical science, ecology, social science, reliability, etc. They can be described as follows: individuals are moving progressively along a finite set of states and observations are made in a time window split into several intervals. At each observation time, the only available information is the number of individuals in each state and the history of each item viewed as a stochastic process is thus lost. The time spent in a given state is unknown. Using a data completion technique, an estimation of the hazard rate in each state based on sojourn times is obtained and an estimation of the survival function is deduced. These methods are studied through simulations and applied to a data set. The simulation study shows that the algorithms involved in the methods converge and are robust.     

Kendall distributions and level sets in bivariate exchangeable survival models

For a given bivariate survival function , we study the relations between the set of the level curves of and the Kendall distribution. Then we characterize the survival models simultaneously admitting a specified Kendall distribution and a specified set of level curves. Attention will be restricted to exchangeable survival models.     

离子束辐照下微生物、植物组织存活模型的研究

从低能离子辐照的能量沉积、质量沉积和电荷交换三要素出发,认为质、荷效应对ＤＮＡ具有间接的辐射保护作用,推导出了一个描述低能离子辐照下生物体的存活率公式,给出了典型曲线,并运用该公式对一些实验资料理行了很好的数据拟合。通过反证法,对质量沉积产物对ＤＮＡ具有损伤作用的假设提出质疑,进一步证明了所建模型之立论基础的可信性。     

Generalized viscoelastic designer functionally graded auxetic materials engineered/tailored for specific task performances

For arbitrary linear Kelvin model viscoelastic constitutive relations, generalized analyses based on collocation, least squares, Lagrangean multipliers, calculus of variation and inverse formulations are presented for determining viscoelastic designer material properties tailored and engineered to be best suited for specific boundary and loading conditions and their time histories. Optimum 3-D anisotropic designer materials, including auxetic viscoelastic functionally graded ones, are studied to minimize thermal stresses, creep buckling, creep rates, deflections, aero- and hydro- dynamic noise and static and dynamic aero-viscoelastic effects while concurrently lowering failure probabilities and extending structural survival times and maximizing or minimizing energy dissipation and its rate. The analyses are formulated for single structural elements as well as the entire structure. Extensions to the entire vehicle that incorporate aerodynamics, stability and control are discussed and the dimensions of computational requirements are estimated.