This paper describes the program module ＂GRAPHS＂ which was developed for data processing in geobotany and ecology fields. The ＂GRAPHS＂ has a simple interface and is integrated into the Microsoft Excel. This allows...This paper describes the program module ＂GRAPHS＂ which was developed for data processing in geobotany and ecology fields. The ＂GRAPHS＂ has a simple interface and is integrated into the Microsoft Excel. This allows users to use all features of Microsoft Excel for storage and preparation data for analysis. Calculation of the most common similarity indexes （Jaccarda. Sorenson, Ohai etc.） and their visualization by using different algorithms of the graph theory or hierarchical cluster analysis allows simplifying and accelerating the process of data analysis in ecology or geobotany and makes it clearer. Also, three ordination methods--PCA （principal components analysis）, CA （correspondence analysis）. NMS （nonmetric multidimensional scaling）-were implemented in the module. The module can be used for vegetation classification, and be used to allocate diagnostic species or to search environmental factors most strongly impact on vegetation. Algorithms of data analysis which were implemented in the module ＂GRAPHS＂ have universal nature so they can be applied in many other fields of science.展开更多
提出一种识别输电断面的快速方法,为有针对性地实施输电断面的保护与控制策略提供前提。首先根据电网实时动态监测系统(Wide area measurement system,WAMS)提供的广域电压相角信息,采用基于离差平方和法的聚类分析方法,将待搜索电网或...提出一种识别输电断面的快速方法,为有针对性地实施输电断面的保护与控制策略提供前提。首先根据电网实时动态监测系统(Wide area measurement system,WAMS)提供的广域电压相角信息,采用基于离差平方和法的聚类分析方法,将待搜索电网或区域电网内母线分为2群;进一步基于图论中极大连通子图的概念,给出了在母线群内搜索电源区或负荷区的图论方法;最后将相邻电源区和负荷区间的联络线簇识别为输电断面。在此基础上,给出了输电断面的层次递进搜索策略及其停止准则,实现各级区域电网间和区域电网内输电断面的搜索。仿真结果验证了所提方法的有效性。展开更多
在分析谱聚类实现思路和已有算法基础上,对规范切判据,最小最大切判据和自动确定聚类数目的谱聚类典型算法进行了研究和应用,通过理论分析算法各自实现机理的联系与区别,讨论它们各自的聚类特点,并利用UCI(University of California,Irv...在分析谱聚类实现思路和已有算法基础上,对规范切判据,最小最大切判据和自动确定聚类数目的谱聚类典型算法进行了研究和应用,通过理论分析算法各自实现机理的联系与区别,讨论它们各自的聚类特点,并利用UCI(University of California,Irvine)机器学习数据集试验对比了三种算法的聚类效果。发现谱聚类算法实现数据聚类的有效性,以及参数及相似度度量对算法性能有很大影响,在此基础上提出了算法用于解决可建模为模式识别的工程问题的可行思路,为工程实践提供了借鉴。展开更多
文摘This paper describes the program module ＂GRAPHS＂ which was developed for data processing in geobotany and ecology fields. The ＂GRAPHS＂ has a simple interface and is integrated into the Microsoft Excel. This allows users to use all features of Microsoft Excel for storage and preparation data for analysis. Calculation of the most common similarity indexes （Jaccarda. Sorenson, Ohai etc.） and their visualization by using different algorithms of the graph theory or hierarchical cluster analysis allows simplifying and accelerating the process of data analysis in ecology or geobotany and makes it clearer. Also, three ordination methods--PCA （principal components analysis）, CA （correspondence analysis）. NMS （nonmetric multidimensional scaling）-were implemented in the module. The module can be used for vegetation classification, and be used to allocate diagnostic species or to search environmental factors most strongly impact on vegetation. Algorithms of data analysis which were implemented in the module ＂GRAPHS＂ have universal nature so they can be applied in many other fields of science.
文摘提出一种识别输电断面的快速方法,为有针对性地实施输电断面的保护与控制策略提供前提。首先根据电网实时动态监测系统(Wide area measurement system,WAMS)提供的广域电压相角信息,采用基于离差平方和法的聚类分析方法,将待搜索电网或区域电网内母线分为2群;进一步基于图论中极大连通子图的概念,给出了在母线群内搜索电源区或负荷区的图论方法;最后将相邻电源区和负荷区间的联络线簇识别为输电断面。在此基础上,给出了输电断面的层次递进搜索策略及其停止准则,实现各级区域电网间和区域电网内输电断面的搜索。仿真结果验证了所提方法的有效性。
文摘在分析谱聚类实现思路和已有算法基础上,对规范切判据,最小最大切判据和自动确定聚类数目的谱聚类典型算法进行了研究和应用,通过理论分析算法各自实现机理的联系与区别,讨论它们各自的聚类特点,并利用UCI(University of California,Irvine)机器学习数据集试验对比了三种算法的聚类效果。发现谱聚类算法实现数据聚类的有效性,以及参数及相似度度量对算法性能有很大影响,在此基础上提出了算法用于解决可建模为模式识别的工程问题的可行思路,为工程实践提供了借鉴。
