On the other hand , they play an important role in the theories of esfimation for regression function . in this paper , we mainly get the large sample properties for partitioning estiona - tion and modified its estimation . for example , we proved their asymptolic normaity under nuture conditions by means of mortingle theory ; we also get their strong consistency for regression function under censored samples ; and finaly we genearzed the result to dependence sample and have strong consistency for the modified partitioning estimation of regression function 因此本论文研究了回归函数基于分割估计及改良基于分割估计的大样本性质，利用鞅的有关理论，在比较自然的条件下，证明了其渐近正态性；首次构造了截尾样本的回归函数基于分割估计及改良基于分割估计，并证明其强相合性；同时把有关结果推广到相依样本下(如混合) ，获得了改良基于分割估计的强相合性及收敛速度。
Based on the basic principle of wavelet analysis , multi - resolution analysis of signal in atomic clock is made . the coefficients of wavelet transform for synthesis atomic time , which are weighed and averaged at different wavelet scales , are obtained . then according to reconstruction theorem , multi - resolution synthesis time scale can be reconstructed . since the signal of atomic clock has been analysed by multi - resolution , we can use wavelet variance at different scales to weight and average the coefficients . thus either the difference in stability of different clock or the varying characteristic of the same clock at the different scale are all considered . finaly , this method is checked by the measured data from national timing serve center of shaanxi astronomy observatory . it confirmed that this technique , which is simple and practicable , is a new method of multi - resolution . from this method , the common characteristic of different clocks can be extracted . the stability of multi - resolution synthesis atomic time scale is obviously superior to that of other methods 本文根据小波分析的基本原理，对原子钟信号进行多分辨率分解，将分解后的小波变换系数进行加权平均，得到不同小波尺度综合原子时的加权平均小波变换系数，然后由小波变换的重构条件，反演综合时间尺度.由于对原子钟信号进行了小波分解，利用不同尺度的小波变换系数的小波方差进行加权平均，这样既考虑不同原子钟在稳定性方面的差异，又顾及同一台原子钟在不同小波尺度的变化特性.最后根据陕西天文台国家授时中心的实测数据对这种方法进行了检验.表明这是一种全新的多分辨率综合方法，这种方法比较简单而切实可行，它能提取各个原子钟的共同特性，多分辨率综合时间尺度的平稳性明显优于其他方法
Finaly , we provide a proposal of constructing foundation class libration of c + + according to semantic of capsule and protocol , as well as the method to construct the domain class derived from the foundation class . it enable us to construct a formal model for real - time system concurrency which is the foundation of generation of code and also to validate system by formal analyse method 从而使得我们能够利用这些元素在模型层建立关于一个具体的实时系统的并发性的形式化的uml模型，并可以利用形式化的方法来分析和验证系统的正确性，同时也对系统运行代码的自动生产奠定了基础。
The heat transfer model of lpg vessel under flame surroundings is set up first , then the heat transfer boundary conditions of lpg vessel are determined according to the heat transfer theory and the heat exchange boundary conditions of lpg vessel of global shape with 4l percellt fill rate is calculated in terms of plgs99 software , finaly the temperafore field distribution of this model is calculated and analyzed by ansys software 在液化气容器的温度场分布计算中，本文首先建立了火焰包围下液化气容器的传热模型，然后根据传热学理论确定了液化气容器瞬态传热的边界条件，并结合plgs99软件计算出了填充量为41的球形液化气压力容器模型的换热边界条件，最后应用ansys计算了该模型的瞬态温度场分布，并进行了分析。
Finaly , according to the requirements of our army for intelligentizing and informationazing management , omkbs is applied in the combat readiness equipment support evaluation system in the thesis . by building commonkads model and omkbs model for the system , the value of knowledge system in military field has been proved in the thesis 最后,针对我军对智能化、信息化管理的需求,将omkbs建模框架应用于装备战备保障评估系统中,对该系统进行了知识建模,表明了知识系统在军事领域中的应用价值。
Discussing on the present development of macroeconomic early - warning , this paper points to the present problems about macroeconomic early - warning ; at the same time , it introduces the theory of macroeconomic early - warning and discusses on the problems , such as , the design of early - warning ( e - w ) indexes , the choice of e - w model , and the decision of e - w deadline . by analyzing two methods : artificial neural network ( ann ) and rough set ( rs ) , we connect the theory of rough set with artificial neural network and build the model of macroeconomic early - warning system based on rough set and artificial neural network ( me - rsann ) . on the basis of it , the paper introduces the basic content about me - rsann and researches the realization of early - warning ways including the structure model of ms - rsann , resolvement of problems during operation of early - warning model , the design of specific early - warning deadline , and so on ; finaly , using statistics data , we carry out quantitative analysis according to the model of me - rsann 本文首先对国内外宏观经济预警进行概述，提出目前在宏观经济预警方面存在的主要问题；并在此基础上阐述宏观经济预警理论体系，讨论了预警指标的设计、预警方法的选择和预警警限的界定等问题；在分析和讨论人工神经网络( ann )预警方法和粗集( rs )理论方法的优势互补后，建立了基于粗集与神经网络相结合宏观经济预警系统模型，详细地论述了基于粗集-神经网络预警方法的基本原理，研究了预警方法的实现，其中包括建立粗集-神经网络预警的结构模式、解决粗集-神经网络预警操作中的问题、提出具体的预警警限的界定方法；最后，利用我国的统计资料，对ann预警和粗集-神经网络相结合预警方法进行了实证分析和比较研究。
However , in this thesis , the writer made a deeper study on this topic , and finaly made a definite concept of the world manufacturing center . she thinks that a production center is to be clustered by certain world manufacturing industries and supported by a higher and deeper industrial chain . this center should be formed on the land of china , and more than half of its products are produced for world market and its quantity of production should be more than 1 / 10 of the world total production 理论意义在于，国内和国外有很多专家和学者对于全球制造中心这一个概念有过相关的论述，并提出了自己的见解，但是，就笔者所了解的文献来看，目前还没有专家和学者对于全球制造中心提出其明确的内涵意义，本文对于全球制造中心的内涵做了一些初步的研究尝试，提出了自己的见解，认为，中国的全球制造中心是指在中国国土上形成的，行业过半产品面向世界市场出口的，而且产销量和产销额都占到国际市场10以上份额的，拥有上游纵深产业链支持的，由若干个这样的全球制造行业聚集形成的生产制造活动中心。
