为提高名词性属性实例差异的识别精度,优化分类算法性能,综合考虑实例的属性和类别特征,提出了一种基于条件概率分布的混合距离度量方法.首先,计算属性间以及属性与类别间条件概率分布的差异;其次,利用互信息对2种差异进行加权组合,得...为提高名词性属性实例差异的识别精度,优化分类算法性能,综合考虑实例的属性和类别特征,提出了一种基于条件概率分布的混合距离度量方法.首先,计算属性间以及属性与类别间条件概率分布的差异;其次,利用互信息对2种差异进行加权组合,得到新的混合距离度量;最后,利用K-近邻算法在20个UCI(University of California Irvine)数据集上进行仿真实验,并将其应用于儿童阑尾炎的诊断和治疗.结果表明:较重叠度量等3种度量方法,本文提出的距离度量方法,显著提高了分类算法的准确率,具有较好的应用前景.展开更多
风向作为风荷载变异性的关键因素之一,在极端风速预测中不容忽视。基于安徽省67个气象站点1980—2020年间的历史日极值风速和风向数据,分别对风速和风向进行了概率分布建模分析。结合最优边际分布,采用3种常用的阿基米德Copula函数建立...风向作为风荷载变异性的关键因素之一,在极端风速预测中不容忽视。基于安徽省67个气象站点1980—2020年间的历史日极值风速和风向数据,分别对风速和风向进行了概率分布建模分析。结合最优边际分布,采用3种常用的阿基米德Copula函数建立了风速和风向的二维联合分布模型。最后结合Copula函数的条件概率理论,给出了全省范围内16个风向上50年重现期下的设计风速区划。研究结果表明,Gumbel分布和三阶von Mises分布适用于全省绝大多数站点风速和风向的边际分布。根据赤池信息准则(Akaike information criterion,AIC)可以发现,对于大多数气象站点的二维联合分布,Frank-Copula函数拟合误差最小。各风向下的极值风速大小存在显著差异,忽略风向的极值风速预测,将会导致计算结果不准确。展开更多
This paper is concerned with the global boundedness problem for a class of stochastic nonlinear systems with matched conditions. The uncertainties in the systems are due to parameter variations and external stochastic...This paper is concerned with the global boundedness problem for a class of stochastic nonlinear systems with matched conditions. The uncertainties in the systems are due to parameter variations and external stochastic disturbance. Only the matched conditions and the possible bound of the uncertainties are demanded. Based on the stochastic Lyapunov stability theory, an explicit controller is constructed in the gradient direction, which renders responses of the closed-loop systems be globally bounded in probability. When the systems degrade to linear systems, the controller becomes linear. Illustrative examples are given to show the effectiveness of the proposed method.展开更多
This paper presents a method to acquire runtime distribution ratio of building air conditioning system under part load condition (part load coefficient of system) through practical energy consumption data. By utilizin...This paper presents a method to acquire runtime distribution ratio of building air conditioning system under part load condition (part load coefficient of system) through practical energy consumption data. By utilizing monthly energy consumption data of the entire year as the analysis object,this paper identifies data distribution,verifies distribution characteristics and analyzes distribution probability density for the issue of running time distribution ratio of air conditioning system in part load zones in the whole operation period,thus providing a basic calculation basis for an overall analysis of energy efficiency of air conditioning system. In view of the general survey of public building energy consumption carried by the government of Chongqing,this paper takes the governmental office building as an example,the part load ratio coefficient corresponding to practical running of air conditioning system of governmental office building in Chongqing is obtained by utilizing the above probability analysis and the solving method of probability density function. By utilizing the ratio coefficient obtained using this method,the part load coefficient with any running ratio of air conditioning system can be obtained according to the requirement of analysis,which can be used in any load ratio for analyzing running energy efficiency of air conditioning system.展开更多
A model of constant probability event is constructed rigorously in event space of PSCEA. It is showed that the numericalbased fusion and the algebraicbased fusion have a consistent result when the weight is regarded a...A model of constant probability event is constructed rigorously in event space of PSCEA. It is showed that the numericalbased fusion and the algebraicbased fusion have a consistent result when the weight is regarded as a constant probability event. From the point of view of algebra, we present a novel similarity measure in product space. Based on the similarity degree, we use a similarity aggregation method to fusion experts' evaluation. We also give a numerical example to illustrate the method.展开更多
文摘为提高名词性属性实例差异的识别精度,优化分类算法性能,综合考虑实例的属性和类别特征,提出了一种基于条件概率分布的混合距离度量方法.首先,计算属性间以及属性与类别间条件概率分布的差异;其次,利用互信息对2种差异进行加权组合,得到新的混合距离度量;最后,利用K-近邻算法在20个UCI(University of California Irvine)数据集上进行仿真实验,并将其应用于儿童阑尾炎的诊断和治疗.结果表明:较重叠度量等3种度量方法,本文提出的距离度量方法,显著提高了分类算法的准确率,具有较好的应用前景.
文摘风向作为风荷载变异性的关键因素之一,在极端风速预测中不容忽视。基于安徽省67个气象站点1980—2020年间的历史日极值风速和风向数据,分别对风速和风向进行了概率分布建模分析。结合最优边际分布,采用3种常用的阿基米德Copula函数建立了风速和风向的二维联合分布模型。最后结合Copula函数的条件概率理论,给出了全省范围内16个风向上50年重现期下的设计风速区划。研究结果表明,Gumbel分布和三阶von Mises分布适用于全省绝大多数站点风速和风向的边际分布。根据赤池信息准则(Akaike information criterion,AIC)可以发现,对于大多数气象站点的二维联合分布,Frank-Copula函数拟合误差最小。各风向下的极值风速大小存在显著差异,忽略风向的极值风速预测,将会导致计算结果不准确。
基金supported by the National Natural Science Foundation of China(61304020)
文摘This paper is concerned with the global boundedness problem for a class of stochastic nonlinear systems with matched conditions. The uncertainties in the systems are due to parameter variations and external stochastic disturbance. Only the matched conditions and the possible bound of the uncertainties are demanded. Based on the stochastic Lyapunov stability theory, an explicit controller is constructed in the gradient direction, which renders responses of the closed-loop systems be globally bounded in probability. When the systems degrade to linear systems, the controller becomes linear. Illustrative examples are given to show the effectiveness of the proposed method.
基金Project(50838009) supported by the National Natural Science Foundation of ChinaProjects(2006BAJ02A09,2006BAJ02A13-4) supported by the National Key Technologies R & D Program of ChinaProject(CSTC,2008AB7110) supported by the Key Technologies R & D Programme of Chongqing,China
文摘This paper presents a method to acquire runtime distribution ratio of building air conditioning system under part load condition (part load coefficient of system) through practical energy consumption data. By utilizing monthly energy consumption data of the entire year as the analysis object,this paper identifies data distribution,verifies distribution characteristics and analyzes distribution probability density for the issue of running time distribution ratio of air conditioning system in part load zones in the whole operation period,thus providing a basic calculation basis for an overall analysis of energy efficiency of air conditioning system. In view of the general survey of public building energy consumption carried by the government of Chongqing,this paper takes the governmental office building as an example,the part load ratio coefficient corresponding to practical running of air conditioning system of governmental office building in Chongqing is obtained by utilizing the above probability analysis and the solving method of probability density function. By utilizing the ratio coefficient obtained using this method,the part load coefficient with any running ratio of air conditioning system can be obtained according to the requirement of analysis,which can be used in any load ratio for analyzing running energy efficiency of air conditioning system.
文摘A model of constant probability event is constructed rigorously in event space of PSCEA. It is showed that the numericalbased fusion and the algebraicbased fusion have a consistent result when the weight is regarded as a constant probability event. From the point of view of algebra, we present a novel similarity measure in product space. Based on the similarity degree, we use a similarity aggregation method to fusion experts' evaluation. We also give a numerical example to illustrate the method.
