查询词典 sufficient statistic

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。

But, very few somebody thinks of,"" of " of civilian battalion economy, enterprise of " civilian battalion is not the legal concept on firm sense and normative meaning, in the file that also did not appear at the party almost, and cannot find the statistical caliber of direct correspondence and data from inside the official statistic of national statistic bureau.

但是，很少有人想到，"民营经济"、"民营企业"并非严格意义和规范意义上的法律概念，也几乎未出现于党的文件之中，并且无法从国家统计局的官方统计中找到直接对应的统计口径与数据。

In 1960's, the famous statist,named Moore, proved the asymptotic normality of one kind of linear order statistic. While Hajek gained the equational condition of the asymptotic normality of rank statistic.

早在二十世纪六十年代，Moore构造了一种较为一般的线性次序统计量，并证明了它的渐近正态性，而Hajek也得出了著名的Hajek基本定理，找出了线性秩统计量渐近性的一个充要条件。

On the basis of direct modelling method by using extended Kalman filter, a new method combining extended Kalman filter with adaptive extended Kalman filter is presented to model gyro drift directly from rotate angle of turntable. Extended Kalman filter has great estimation error, what is more, may be divergent if the model and noise statistic are not accurate or wrong. The new method solves the problem of general extended Kalman filter by using fictious noise with timevarying statistic to compensate the linearized model error of extended Kalman filter. The parameter estimators of single-axis and double-axis servo test are designed from it.

本文深入研究了根据伺服试验数据进行陀螺仪漂移误差模型建模的方法，在应用广义Kalman滤波器直接建模的方法基础上，针对广义Kalman滤波器对于不精确或错误的模型和噪声统计，估计误差较大，甚至滤波发散的缺陷，借助于用带时变噪声统计的虚拟噪声补偿广义Kalman滤波的线性化模型误差的新思想，提出了用广义Kalman滤波和自适应广义Kalman滤波算法相结合的方法由转台的转角数据直接建模陀螺仪漂移误差模型，并设计了陀螺仪单轴伺服和双轴伺服测试时参数估计的滤波器。

At first, the characteristics of probability distribution and frequency spectrum of noise inside measuring signal of feedback elements are discussed. Statistic filter method is come up with for noise signal of being Gauss distribution. It is analyzed reason that statistic filter method is superior to arithmetic average filter method.

分析了反馈元件测量信号中噪声的概率分布特性和频谱特性，对呈正态分布的噪声信号提出了统计滤波法，分析了该滤波法优于算术平均滤波法的原因。

This paper proves that Ordered Statistic with Greatest Option CFAR and Ordered Statistic with Smallest Option CFAR are also of CFAR under homeomorphic independent K-distributed clutter background when the shape parameter is known. The performance of two detectors is analyzed against uniform clutter background, clutter edges and multiple target environments.

摘要该文证明了形状因子已知条件下OSGO-CFAR和OSSO-CFAR检测器在均匀统计独立的K分布杂波背景下具有恒虚警性能，分析了两种检测器在均匀杂波背景、杂波边缘和存在强干扰目标情况下的检测性能。

This is station of saying famous careless root statistic, not famous for instance: Hubble-bubble stationmaster statistic, also be the statistical system that leaving.

这是说的知名的草根统计站，不知名的比如：泡泡站长统计，也是在下的统计系统。

This paper presents a self-adaptive load balancing algorithm, in which each peer creates a local load distribution view using a passive load statistic method and a local file requested view using file requested statistic method. When load imbalance exists in the system, the heavy loaded peer will make the logical links which point to itself to point to a light loaded peer in its local load distribution view, with the indegree of the heavy loaded peer decreasing and the indegree of the light loaded peer increasing, the load imbalance magnitude will decrease. When the request load of the heavy loaded peer is high, the peer will use its local file request view to get the popular file and cache the file to corresponding target peer.

提出一种自适应负载均衡方法，方法采用一种被动式结点负载统计方法生成局部负载视图；采用一种文件访问统计方法生成局部文件访问视图；当系统内结点负载存在差异，重载结点把指向自身的逻辑链路迁移至指向局部负载视图中的轻载结点，通过减小重载结点入度和增加轻载结点入度来减小结点间负载差异；当结点的请求负载较高时，通过局部文件访问视图计算需要缓存的热点文件及目标结点，降低承载热点文件的结点的请求负载。

The research on the statistic analysis and the prediction model of road accident is an important part of the system. Through the statistic analysis of various factors which may lead to the accident, it reveals the intrinsical rule and causes of accidents, based on which we can predict the future trend of accidents in order to provide reliable information for effective accident prevention.

道路交通事故统计分析及预测模型研究是该系统的一个子课题，其主要是通过对诱发交通事故的各种因素进行统计分析，揭示交通事故发生的规律及原因，在此基础上对事故未来的发生趋势作出预测，为事故的有效预防提供可靠信息。

By 〓,weconstruct testing statistic of change point and derive the limit distribution of thetesting statistic.

Gombay and 〓(1998)在上述结果的基础上，讨论了变点的估计问题，得到了变点估计的渐近分布，并且讨论了变点估计的相合性。

Yet, Edith says, there was always money in the bank.

但是Edith说，她的银行户口里面总是有钱。

We have created a series of files that resolve this problem--one for each day of the 40 days of purpose.

我们制作一系列的档案去解决这问题---一个为标竿40天中的每一天。