It is well acknowledged that quality of software has a higher priority than the performance and functions of software. One of the ways to get high-quality software is to get more efficient software-testing techniques....It is well acknowledged that quality of software has a higher priority than the performance and functions of software. One of the ways to get high-quality software is to get more efficient software-testing techniques. Theory and technology of software quality assurance are an important part of software developing methodology and software engineering. Software testing plays a key role in software quality assurance. The purpose of the essay is to search for new software testing method and to solve some problems in testing of object-oriented program. We also try to amend some deficiency in the traditional test method for structured programs. By the idea of program slicing, we can disassemble the source code of a program into several slices following certain rules. Instead of testing the whole program, we can test these slices. We can also guarantee the equivalence of the two ways. Testing on the base of program slicing has several advantages than the one simply using data flow analysis and control flow analysis. The first, because a program equals to the union of its slices, to test all of the slices makes a complete test of the program, and to test each slice which is related to the interested variables is actually a complete test of the requirement test. Then we solve the problem of sufficiency in traditional structured program testing and object-oriented program testing as well. The second, program slicing technique can be applied to the testing of both structured programs and object-oriented ones.展开更多
贝叶斯网络由于其强大的不确定性推理能力和因果可表示性越来越受到研究者的关注。从数据中学习一个贝叶斯网络结构被称为NP-hard问题。其中,针对K2算法强依赖于变量拓扑序的问题,提出了一种组合变量邻居集和v-结构信息的K2改进学习方法...贝叶斯网络由于其强大的不确定性推理能力和因果可表示性越来越受到研究者的关注。从数据中学习一个贝叶斯网络结构被称为NP-hard问题。其中,针对K2算法强依赖于变量拓扑序的问题,提出了一种组合变量邻居集和v-结构信息的K2改进学习方法TSK2(Two-Step Search Strategy of K2)。该方法有效减小了序空间搜索规模,同时避免了过早陷入局部最优。具体而言,该方法在约束算法定向规则的启示下,借助识别的v-结构和邻居集信息可靠调整汇点的邻居在序中的位置;其次,在贝网基本组成结构的启发下,借助变量邻居集信息,通过执行顺连、分连、汇连3个基本结构的搜索,准确修正父节点与子节点的序位置,获得最优序列。实验结果表明,在Asia和Alarm网络数据集上,与对比方法相比,所提算法的准确率得到显著提升,可以获得更准确的网络结构。展开更多
概念漂移是流数据的主要特征之一,如何检测概念漂移的发生以及调整预测模型去适应概念漂移现象备受研究者的关注.目前有关概念漂移的大多数算法仅仅针对单一类型的概念漂移检测,并且需限制输入数据服从某一分布,所以在检测多种类型概念...概念漂移是流数据的主要特征之一,如何检测概念漂移的发生以及调整预测模型去适应概念漂移现象备受研究者的关注.目前有关概念漂移的大多数算法仅仅针对单一类型的概念漂移检测,并且需限制输入数据服从某一分布,所以在检测多种类型概念漂移时效果不理想.提出一种在线集成自适应算法(KSHPR),在自适应随机森林(Adaptive Random Forests,ARF)算法和流随机补丁(Streaming Random Patch,SRP)算法的基础上进行优化改进,采用非参数检验与滑动窗口相结合的策略进行概念漂移检测,降低窗口平均值对算法性能的影响,并以此为基础建立四个基学习者的集成学习模型,根据基学习者预测准确率,动态分配权值,有效解决流式数据中学习模型精度低的问题.实验证明,提出的算法在真实数据集和合成数据集中均表现优良,与其他算法相比,该算法的稳定性、分类准确性与多类型概念漂移适应能力均有所提升.展开更多
RoboCup is a particularly good domain for studying multi-agent systems. A wide variety of MAS issues can be studied in robotic soccer, in which the theory, algorithm and architecture of agent system can be evaluated. ...RoboCup is a particularly good domain for studying multi-agent systems. A wide variety of MAS issues can be studied in robotic soccer, in which the theory, algorithm and architecture of agent system can be evaluated. Because of the inherent complexity of MAS, there are many interests in using machine learning techniques to handle it. This paper investigates and discusses the machine-learning techniques used in RoboCup. The background is firstly presented and the application of machine learning in RoboCup is lately demonstrated with some top simulation teams. The machine-learning system in NDSocTeam is also introduced. Finally some open issues in this field are pointed out.展开更多
1.引言近年来,国内外对并行文件系统做了很多的研究,著名的有:一些商用的并行文件系统,如Intel为iPSC/2和iPSC/860而设计的CFS,以及IntelParagon的并行文件系统PPFS。还有HPF的运行支撑系统PASSION。研究性的则有PIOUS等。它们通常是针...1.引言近年来,国内外对并行文件系统做了很多的研究,著名的有:一些商用的并行文件系统,如Intel为iPSC/2和iPSC/860而设计的CFS,以及IntelParagon的并行文件系统PPFS。还有HPF的运行支撑系统PASSION。研究性的则有PIOUS等。它们通常是针对并行机设计的,应用范围较为狭窄。随着对NOW(Network of Workstations,工作站网络)的研究深入。展开更多
文摘It is well acknowledged that quality of software has a higher priority than the performance and functions of software. One of the ways to get high-quality software is to get more efficient software-testing techniques. Theory and technology of software quality assurance are an important part of software developing methodology and software engineering. Software testing plays a key role in software quality assurance. The purpose of the essay is to search for new software testing method and to solve some problems in testing of object-oriented program. We also try to amend some deficiency in the traditional test method for structured programs. By the idea of program slicing, we can disassemble the source code of a program into several slices following certain rules. Instead of testing the whole program, we can test these slices. We can also guarantee the equivalence of the two ways. Testing on the base of program slicing has several advantages than the one simply using data flow analysis and control flow analysis. The first, because a program equals to the union of its slices, to test all of the slices makes a complete test of the program, and to test each slice which is related to the interested variables is actually a complete test of the requirement test. Then we solve the problem of sufficiency in traditional structured program testing and object-oriented program testing as well. The second, program slicing technique can be applied to the testing of both structured programs and object-oriented ones.
文摘贝叶斯网络由于其强大的不确定性推理能力和因果可表示性越来越受到研究者的关注。从数据中学习一个贝叶斯网络结构被称为NP-hard问题。其中,针对K2算法强依赖于变量拓扑序的问题,提出了一种组合变量邻居集和v-结构信息的K2改进学习方法TSK2(Two-Step Search Strategy of K2)。该方法有效减小了序空间搜索规模,同时避免了过早陷入局部最优。具体而言,该方法在约束算法定向规则的启示下,借助识别的v-结构和邻居集信息可靠调整汇点的邻居在序中的位置;其次,在贝网基本组成结构的启发下,借助变量邻居集信息,通过执行顺连、分连、汇连3个基本结构的搜索,准确修正父节点与子节点的序位置,获得最优序列。实验结果表明,在Asia和Alarm网络数据集上,与对比方法相比,所提算法的准确率得到显著提升,可以获得更准确的网络结构。
文摘概念漂移是流数据的主要特征之一,如何检测概念漂移的发生以及调整预测模型去适应概念漂移现象备受研究者的关注.目前有关概念漂移的大多数算法仅仅针对单一类型的概念漂移检测,并且需限制输入数据服从某一分布,所以在检测多种类型概念漂移时效果不理想.提出一种在线集成自适应算法(KSHPR),在自适应随机森林(Adaptive Random Forests,ARF)算法和流随机补丁(Streaming Random Patch,SRP)算法的基础上进行优化改进,采用非参数检验与滑动窗口相结合的策略进行概念漂移检测,降低窗口平均值对算法性能的影响,并以此为基础建立四个基学习者的集成学习模型,根据基学习者预测准确率,动态分配权值,有效解决流式数据中学习模型精度低的问题.实验证明,提出的算法在真实数据集和合成数据集中均表现优良,与其他算法相比,该算法的稳定性、分类准确性与多类型概念漂移适应能力均有所提升.
文摘RoboCup is a particularly good domain for studying multi-agent systems. A wide variety of MAS issues can be studied in robotic soccer, in which the theory, algorithm and architecture of agent system can be evaluated. Because of the inherent complexity of MAS, there are many interests in using machine learning techniques to handle it. This paper investigates and discusses the machine-learning techniques used in RoboCup. The background is firstly presented and the application of machine learning in RoboCup is lately demonstrated with some top simulation teams. The machine-learning system in NDSocTeam is also introduced. Finally some open issues in this field are pointed out.
文摘1.引言近年来,国内外对并行文件系统做了很多的研究,著名的有:一些商用的并行文件系统,如Intel为iPSC/2和iPSC/860而设计的CFS,以及IntelParagon的并行文件系统PPFS。还有HPF的运行支撑系统PASSION。研究性的则有PIOUS等。它们通常是针对并行机设计的,应用范围较为狭窄。随着对NOW(Network of Workstations,工作站网络)的研究深入。