近似度

Finishing calculation of mean value, standard deviation, skewness, kurtosis of Beta distribution.(2) Fitting parameters of many kinds of typical distribution and using residual deviation to evaluate fitting precision.(3) Using Beta distribution as an agreed indication distribution applied to many kinds of practical photoelectric measurement distributions.(4) Deriving theory formula of Bayes point estimation about Beta distribution parameters and mean value and standard deviation on the condition of mean square error loss function and supposed the prior distribution is uniform distribution.(5) Generating MCMC sample from post distribution by the method of Gibbs sample algorithm. Calculating bayes point estimation from sample on the condition of mean square error loss function. Calculating confidence interval by an approximate method to complete interval estimation.

本文的主要工作有：(1)解决了Beta分布参数a和b的精确计算以及均值、标准差、偏度、峰度的计算问题；(2)拟合出10余种典型分布的Beta分布的两个参数，并且采用剩余标准差评价该Beta分布的拟合精度；(3)对多种典型的光学与光电测量系统的测量分布进行了Beta分布统示表示；(4)在假设先验分布为均匀分布前提下，得到参数a和b以及均值μ和标准差σ在均方误差损失函数下的贝叶斯点估计理论计算公式；(5)利用直接抽样的Gibbs抽样算法，从后验分布中产生MCMC样本，从样本直接计算均方误差损失函数下的贝叶斯点估计，并使用一种近似方法计算其置信区间，完成区间估计。

Fig 1 illustrate another general phenomenon, the amorphous regions in partially crystalline polymers also assume a glassy state, Tg being independent of degree of crystallinity the a first approximation .

从一级近似来讲。好与结晶度无关，以认为是玻璃态。从一级近似来讲。

Among used machine learning methods, the gradient descent method is widely used to train various classifiers, such as Back-propagation neural network and linear text classifier. However, the gradient descent method is easily trapped into a local minimum and slowly converges. Thus, this study presents a gradient forecasting search method based on prediction methods to enhance the performance of the gradient descent method in order to develop a more efficient and precise machine learning method for Web mining.However, a prediction method with few sample data items and precise forecasting ability is a key issue to the gradient forecasting search method. Applying statistic-based prediction methods to implement GFSM is unsuitable because they require a large number of data items to model a prediction model. In the contrast with statistic-based prediction methods, GM(1,1) grey prediction model does not need a large number of data items to build a prediction model, and it has low computational load. However, the original GM(1,1) grey prediction model uses a mathematical hypothesis and approximation to transform a continuous differential equation into a discrete difference equation in order to model a forecasting model.

其中梯度法是一个最常被使用来实现机器学习的方法之一，然而梯度法具有学习速度慢以及容易陷入局部最佳解的缺点，因此，本研究提出一个梯度预测搜寻法则(gradient forecasting search method， GFSM)来改善传统梯度法的缺点，用来提升一些以梯度学习法则为基础的分类器在资讯探勘上的效率与正确性；而一个所需资料量少、计算复杂度低且精确的预测模型是梯度预测搜寻法能否有效进行最佳解搜寻之关键因素，传统统计为基础之预测方法的缺点是需要较大量的数据进行预测，因此计算复杂度高，灰色预测模型具有建模资料少且计算复杂度低等优点，然而灰色预测理论以连续之微分方程式为基础，并且透过一些数学上的假设与近似，将连续之微分方程式转换成离散之差分方程式来对离散型资料进行建模及预测，这样的作法不尽合理，且缺乏数学理论上的完备性，因为在转换过程中已经造成建模上的误差，且建模过程仅考虑相邻的两个资料点关系，无法正确反应数列未来的变化趋势。

Using the infinite product of unevenly distributed probability space with potential of 3, the concept of formula truth degree was introduced into the Gdel 3-valued propositional logic system. Moreover, it was proved that in the 3-value (1/4, 1/2, 1/4), measure the set of truth of degree of propositions was dense in [0, 1], and a general expression of truth degree was obtained, providing a basis for further developing approximate reasoning of 3-valued propositional logic system.

利用势为3的非均匀概率空间的无穷乘积，在Gdel三值命题逻辑系统中引入公式的真度概念，在三值逻辑(四分之一，二分之一，四分之一)测度下证明G3中全体公式的真度值之集在[0， 1]上是稠密的，并给出公式真度的表达通式，为进一步在三值命题逻辑系统中展开近似推理奠定基础。

Main contents and conclusions are: 1 The ionization current was measured continuously by a charge-integration method with a computer after the introduction of radon into the chamber.

主要内容及结论：1充氡后利用计算机连续进行电离电流累积测量，得到了精细的电流随时间的变化曲线，可清楚看出氡及其子体的生成衰变造成的电离电流的生长—稳定—衰减过程，与理论计算值进行了比较，由电离电流衰减速度的变化得出氡吸附作用的影响，并测量到电离室冲洗后氡的反扩散现象；2由已知活度的标准源定出相对测量的刻度系数K值，K=24.62 Bq／pA，对同一个监测源以半年为周期进行多次测量后得出稳定性符合要求；3通过测量电离电流和气压的关系曲线以及不同气压下的饱和曲线，确定了常规的工作条件；4测量了盐酸清洗电离室内表面降低本底的效果；5不确定度分析表明相对测氡的扩展不确定度(k=2)小于5％，由本底涨落得出探测下限为0.5 Bq；6理论上对壁损失修正进行了计算，实验上保持电离室工作在饱和区的情况下，测量了不同气压下电离电流的变化，从中得出了壁损失随气压变化的规律，可以看出，电离电流的实测值和计算值随气压变化规律基本一致；当气压大于125 kPa时，端效应等其它因素造成的影响不大于2％，因而在此工作条件下，测量电离电流后通过计算修正壁损失可以实现近似的绝对测氡，提出了进一步提高绝对测量准确度的方法。

The polynomial function bases are used to form the approximately matching function. Moreover, the discrete values of grey image is transformed to polinomial function that are exact fit at the considered pixel locations. As a result, The mathematical operations performed such as derivation, integration, and gradient evaluation also give out the polynomial functions, therefore, the system of undetermined coefficient equations is derived to be a system of linear algebraic equations.

我们把立体匹配过程看作泛函的极小化过程，因此选用变分问题近似解法的Ritz法构造数值算法，以多项式函数为基构作匹配函数的近似解，进而又利用多项式过样值点的拟合将离散的灰度图象变换为多项式函数，因此计算中所需的求导，积分，及梯度运算的结果仍然是多项式函数，最终将待定系数方程组化为线性代数方程组。

The notion of weak fuzzy similarity relations,a generalization of fuzzy similarity relations,is used to represent indiscernibility of elements.

为此提出用模糊相似关系和弱模糊相似关系的概念来代替等价关系，同时引入了弱模糊相似关系的相似度概念和相似类概念，定义了基于弱模糊相似关系的广义粗糙集及标准的广义粗糙集上下近似；研究了两对上下近似算子的性质，讨论了3种粗糙隶属函数的性质并验证了它们的性质。

Then, according to phase space reconstruct technique from one-dimensional and multi-dimensional time series, the quantitative criterion and rule of system chaos which combine neural network, analyses, computations and sort are conducted on EEG signals of five kinds of human consciousness activities. The author finds that statistic results of the determinism computation in time series indicate that chaos characteristic may lie in human consciousness activities, and central tendency measure is consistent with phase graph, so it may become one division way of attractor; approximate entropy of different subjects exists discrepancy; correlation dimension and Lyapunov exponent of different consciousness activities indicate that attractors of human activities are all fractional; nonlinear quantitative criterion and rule, which unites neural network, can distinguish different styles of consciousness activity in principle, and the result of sortation takes on that consciousness activity type of arithmetic is distinguished more easily than that of abstract.

采用一维与多维时间序列相空间重构技术和系统混沌的定量判据准则，结合神经网络，对五种思维作业方式的脑电信号进行了分析、计算与分类，发现确定性计算的统计结果表明人类的意识活动中可能蕴含混沌特性，中心趋向测量(Central Tendency Measure，CTM)与相图吻合较好，可作为EEG吸引子的区分方法之一；功率谱分析反映出单个受试者的各种意识形态的差异很小，但是不同意识活动谱中的活跃频段还是略有差异的；个体之间的近似熵存在差异，同种状态下近似熵大的人可能具有更好的创新性；关联维数和Lyapunov指数的计算结果表明大脑的运动落在具有分维的奇怪吸引子上；非线性量化方法与神经网络相结合，可对人脑思维活动进行较好的分类，并且从分类结果来看数理计算类思维活动较抽象类思维活动有着更好的区分度。

Formally,the process is considered as fuzzed of the lower and upper approximation sets. Finally, an example shows that membership degrees more visualy and reasonably describe that objects belong to the lower and upper approximation of the set.

实例分析说明在不完备信息系统中，考察对象对于上、下近似的隶属度能更直观、合理地反应对象隶属于某一集合的上、下近似的情况。

An arbitrary shaped line current on a horizontal plane is divided into many small line segments and approximated by Hertzian diploes, then both the near and far EM-field of the line curennt is approximated by the summation of the complete near and far fields of Hertz dipoles.

水平面上任意形状线电流分割为许多小段之电流，每一小段之电流近似为赫芝偶极，则在三度空间任意座标平面上无论远近之电磁场即可以赫芝偶极之远近完整电磁场加总合成近似之。

We have no common name for a mime of Sophron or Xenarchus and a Socratic Conversation; and we should still be without one even if the imitation in the two instances were in trimeters or elegiacs or some other kind of verse--though it is the way with people to tack on 'poet' to the name of a metre, and talk of elegiac-poets and epic-poets, thinking that they call them poets not by reason of the imitative nature of their work, but indiscriminately by reason of the metre they write in.

索夫农 、森那库斯和苏格拉底式的对话采用的模仿没有一个公共的名称；三音步诗、挽歌体或其他类型的诗的模仿也没有——人们把&诗人&这一名词和格律名称结合到一起，称之为挽歌体诗人或者史诗诗人，他们被称为诗人，似乎只是因为遵守格律写作，而非他们作品的模仿本质。

The relationship between communicative competence and grammar teaching should be that of the ends and the means.

交际能力和语法的关系应该是目标与途径的关系。