样本问题

On the basis of analysis of existing problems in application of common comprehensive optimization models, the characteristics of optimization are combined with the principle of Artificial Neural Network organically. The ANN analysis situs structure is established and the learning algorism for the ANN is designed. The methods of generating training samples and alternative optimization are put forward.

在分析水利水电工程常用综合优选模型应用中存在问题的基础上，将综合优选问题的特点和人工神经网络原理有机结合，建立了方案综合优选的人工神经网络拓扑结构；设计了相应的网络学习算法；提出了生成训练样本和方案优选方法。

Some methods are discussed to speed up the K-MEANS clustering algorithm with Euclidean and Mahalanobis distance norm.

同时，归纳并整理了一整套基于欧氏距离和马氏距离的K均值快速算法，为了解决传统基于距离函数算法的样本顺序相关问题，提出了针对两类问题的顺序无关聚类算法。

Constructed by standard binary classes support vector machine, present multiclass SVMs are usually very slow to be trained. When a large number of categories of data are to be classified, the training work could be very difficult. By extending the hypersphere one-class SVM to a hypersphere multiclass SVM, we build a fast training classifier HSOC-SVM. Its training speed is higher than that of the present multiclass classifiers, because each category data trains only one HSOC-SVM.

目前的多类分类器大多是经二分类器组合而成的，存在训练速度较慢的问题，在分类类别多的时候，会遇到很大困难，超球体多类支持向量机将超球体单类支持向量机扩展到多类问题，由于每类样本只参与一个超球体支持向量机的训练，因此，这是一种直接多类分类器，训练效率明显提高。

However, in regression model if there are features that are not completely relative or even completely irrelative to the problem, the difference of features' correlative degree to the problem becomes large and it may affect the performance of support vector regression machine.

基于统计学习理论的支持向量回归机有比较好的泛化能力，然而当样本含有与该问题不完全相关甚至完全无关的特征时，会使得各个特征对问题的相关程度差异很大，从而使得支持向量回归机的效果受到影响。

It's pointed out that soft sensor will play more important role in field bus network control system. According to this, a crude solution was addressed. At last, we point out that soft sensor technology must used to build an information warehouse for the whole enterprise, combined with data fusion, data warehouse, data rectification and other related technologies. The main contributions of this dissertation are as follows: The background, requirement, and application situation of soft sensors are expounded, the theories, methods and skills of soft sensing technology are analyzed, and the fruits and problems in current soft sensor technologies are summarized. Some new methods of soft sensor is proposed: A principal component analysis-based secondary variable selection method are proposed; A new conception which modeling data should have gross error detection is addressed, and then a cluster analysis-based modeling data gross error detection method is given.

本文的主要贡献有：对软测量技术根据实践的要求进行了一定的理论研究，针对具体问题提出了新的方法：讨论了辅助变量选择问题，研究了基于主元分析的辅助变量选择方法，该方法克服了传统方法只能利用数学模型产生的仿真数据进行最优辅助变量选择的缺点，可以根据历史数据进行辅助变量选择；提出了建模数据显著误差侦破的概念，指出传统的显著误差侦破研究的是已知过程数学模型的情况，而建模时数学模型是未知的，但是直接来自现场的数据并不能保证不含显著误差，并用基于聚类分析的方法解决了该问题，该方法利用聚类分析原理，直接面对过程数据，不需以过程模型为基础，在此基础上给出了软测量建模过程中样本数据的处理方法。

Hierarchical linear modeling is the most frequently used technique for analyzing multilevel data. In this paper, the twelve key issues in multilevel research and/or using HLM software were reviewed, and provided their exact meanings and feasible solutions. The 12 key issues included intraclass correlation coefficient, sample sizes, formation of organizational constructs, centering problems, fixed effect and random effect, estimation methods, goodness of fit index, explained variation, multicollinearity, robust standard error, level one variance equation, and empirical Bayes estimates.

本文针对阶层线性模式在多层次研究上有关於技术、测量与方法论所遭遇的问题，进行文献整理并讨论其实务意涵，并依据使用时机、抽样议题、资料汇总、分析方法、与模式评估与估算等五方面整理出十二项重要议题与解决之道，包括：组内相关系数的意义与判断准则、多层次研究各层的样本数问题、变数聚合成组织层次问题、中心化的意义、固定效果与随机效果的设定、估计法的选择、适配度的比较、解释变异量的计算、多元共线性的分析、强韧性标准误、第一层误差项变异数与实证贝氏估计值的运用。

A questionnaire was developed to collect data from samples, which were volunteers of correction institutions of Ministry of Justice in southern Taiwan, including Chayi, Tainan, Kaohsiung, and Pingtong Reformatories.

本研究为避免因区域性倾向问题及样本重覆施测所造成的误差，决定在嘉义、台南、高雄及屏东地区，抽取收容人特性及志工服务项目性质相似之机关各一间，即嘉义监狱、台南监狱、高雄监狱和屏东监狱，并采全面施测方式取得研究样本。

We will further carry out stochastic comparison of spacings of record values from one sample and two samples.

那么，我们这里将研究基于单样本或者双样本的记录值间隔的随机比较问题。

The dissertation recommends some kinds of methods and measures: such as adjusting the network configuration, connecting value and threshold value by the total value of all stylebooks, adding part of the adjusting value of the last time to the current adjusting value, transforming the stylebooks to standard value, optimizing activation function, appending threshold value to the putout of the nerve cell and so on. Simultaneity, the author brings forward a new method to optimize the model of artificial neural network . It is using the automatically adaptive genetic algorithm to make the network configuration, connecting value and threshold value of artificial neural network better. The method can make the model better and improve the simulating effect and forecasting precision.Genetic algorithm is an arithmetic based on evolution and genetics used to search the optimization .

本文针对人工神经网络应用中存在的上述问题，介绍了各种改进方法与措施：如用所有样本的总效果对网络权值矩阵和阈值向量进行调整、调整量中加入动量项、标准化训练样本数据、优化激励函数以及给神经元的输出值添加偏置量等；同时还提出了一种新的优化人工神经网络模型的方法，即采用自适应遗传算法对人工神经网络模型的网络结构和权值阈值进行全局优化搜索，以提高大坝安全监测人工神经网络模型的拟合成果和预测精度。

We successively pose a sampling method for robust fuzzy convex optimization, with the sampled problem being polynomially solvable.

对这类问题，本文导出了其确定性等价最优化问题，分析了其可行集的性质，接着对鲁棒模糊凸最优化问题给出了一种样本方法。

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。