样本问题

Firstly, the sectional data from 27 companies in communications and transportation industry, which are listed before 1999, are linearly regressed. We find their capital structure and firm performance are irrelated significantly. Then the section data are quadricly regressed, but the result is also disillusionary, which indicates that the capital structure of listed companies has not the best range only with the sectional data from 1999 to 2003. Finally, in allusion to the problems in the securities market and illogicality of the capital structure, we put forward some methods to optimize the capital structure for listed companies from four facts.

文章首先对交通运输仓储行业1999年以前上市的27家公司的截面数据进行线性回归，发现在分析时段1999年至2003年，样本公司的净资产收益率和单位经济增加值与资产负债率之间线性不相关；然后文章又对该行业上市公司的截面数据进行二次函数回归，同样也没有得到令人满意的结果，这说明仅就1999年至2003年五年的截面数据而言，上市公司资本结构不存在最优区间；最后，文章对我国证券市场存在的问题以及该行业上市公司资本结构的不合理性提出了改进的设想，并从四个方面提出了优化我国上市公司资本结构的建议。

After covering, the gravity centers of covers are recalculated to achieve better effect. The iterativeness of samples is not needed. Therefore, the problems, such as the choice of initial sample and the clustering speed, are solved, which can not be solved with other algorithms.

该算法可以在覆盖后求重心，不断调整所做的覆盖，对大量的样本数据不需要迭代，因而解决了其他聚类算法难以解决的问题：初值的选择和聚类速度。

Research on the related attribute of evidence, influence and method of dealing with is carried out. Result of experiment demonstrates that the method can lead to the reliable decision effectively. Aiming at conflicting problem of decision making, fusion strategy of BPA reassign is adopted in coordinator of GDSS. Aiming at contradictory of decision-making, the process of decision-making is divided into two stages that are learning and decision-making. Each Agent is trained by proper information given by training example to eliminate contradictories among decision-making Agents and correct mistakes that ensure Agents cooperate effectively in the stage of cooperating decision.(3) In order to overcome the flaw of current GDSS that can only support static problem solving, adopting the method that combing Multi-Decision-Making-Agents Reinforcement Learning and the metempirical knowledge learning, the paper puts forward Reinforcement Learning and optimization under dynamic environment.

提出采用证据理论的证据合成作为协调器中决策Agent的融合策略，对群体决策支持系统的协调器进行决策融合过程中可能出现的情况进行了详细的分析，研究了证据的相关性及其在决策融合过程中的影响，提出了相关证据的处理方法，实验结果表明，该方法可有效保证决策结果的可靠性；针对决策过程中的冲突问题，提出协调器采用基本信任指派二次分配的协调策略；针对决策过程中的矛盾，提出将决策过程分为学习和决策两个阶段，各决策者Agent根据有限的训练样本所给出的正确信息进行训练，消除决策者Agent之间的矛盾和修正决策者Agent自身的错误，从而保证合作决策阶段实现有效的合作决策。

This research has proved that an artificial neural network is capable to differentiate the reactional enhancement of malignant intracranial gliomas in postoperative MR images and the enhancement caused by tumor residual. As a innovative differentiation approach, it requires further studies and researches to determine its applicability in practice, including the increase of training samples, image description standardization, effects to doctoral diagnosis, as well as parameter optimization. With improvements in these elements, the performance of the network will be optimized.

本研究成功地证明了人工神经网络能够鉴别颅内恶性胶质瘤术后磁共振图像的反应性强化与肿瘤残存所致的强化，为解决二者的鉴别这一复杂问题提供了一个新的方法，但是本研究只是初步的应用探讨，能否在实际中应用，还有待于进一步的研究，如训练样本的增加、图像描述的规范化、对医师诊断影响程度、参数的优化等，网络的性能会不断完善和提高。

This research has proved that an artificial neural network is capable to differentiate the reactional enhancement of malignant intracranial gliomas in postoperative MR images and the enhancement caused by tumor residual. As a innovative differentiation approach, it requires further studies and researches to determine its applicability in practice, including the increase of training samples, image de脚本ion standardization, effects to doctoral diagnosis, as well as parameter optimization. With improvements in these elements, the performance of the network will be optimized.

本研究成功地证明了人工神经网络能够鉴别颅内恶性胶质瘤术后磁共振图像的反应性强化与肿瘤残存所致的强化，为解决二者的鉴别这一复杂问题提供了一个新的方法，但是本研究只是初步的应用探讨，能否在实际中应用，还有待于进一步的研究，如训练样本的增加、图像描述的规范化、对医师诊断影响程度、参数的优化等，网络的性能会不断完善和提高。

The performance of a disturbance covariance matrix is degradated in space time adaptive processing when it is estimated with secondary data contaminated by targetlike signals.

针对空时自适应处理中样本协方差矩阵受干扰目标污染时检测性能下降的问题，提出了一种基于知识的空时自适应处理（knowledge aided space time adaptive processing， KASTAP）方法。

By getting the FFT angular feature parameters from outline image of tracheid,a lotof common angle of cell walls can be obtained.This method can be used to describing theshape of tracheid and reduced tracheid's feature model,which can't be achieved by traditionalobservation method.

通过管胞轮廓图的FFT角度特征参数提取，还原出大量样本数统计意义上的细胞壁角度，解决了管胞的形状描述问题，还原出管胞的形态模型，这也是传统的观察法无法做到的。

There have some difficulties in the applications of Neural Networks, the convergence for large scale problem is very slowly, the maturity of samples often can not be guaranteed, the calculating results usually trustless because the bad capability of extrapolation of Neural Networks.

人工神经网络模型在实际应用中存在一些困难，包括对于较大规模的问题学习过程缓慢，样本的完备性难以保证，由于人工神经网络的外推计算能力差而导致计算结果的可靠性降低等。

We examine this question using matched primary and seco ndary data on a sample of U.S. manufacturing firms.

我们研究这个问题，通过对美国制造业的一个样本的原始数据和第二手数据进行匹配。

The question we need to ask about the storing of DNA samples is the same as the one bedevilling our approach to identity cards, and stiff anti-terrorism legislation.

这个问题我们要问，对于存放的DNA样本是一样的一个困扰我们的做法，以领取身份证，并在激烈的反恐怖主义立法。

And Pharaoh spoke to Joseph, saying, Your father and your brothers have come to you.

47：5 法老对约瑟说，你父亲和你弟兄们到你这里来了。

Additionally, the approximate flattening of surface strip using lines linking midpoints on perpendicular lines between geodesic curves and the unconditional extreme value method are discussed.

提出了用测地线方程、曲面上两点间短程线来计算膜结构曲面测地线的方法，同时，采用测地线间垂线的中点连线和用无约束极值法进行空间条状曲面近似展开的分析。