For imbalanced datasets, the focus of classification is to identify samples of the minority class. The performance of current data mining algorithms is not good enough for processing imbalanced datasets. The synthetic...For imbalanced datasets, the focus of classification is to identify samples of the minority class. The performance of current data mining algorithms is not good enough for processing imbalanced datasets. The synthetic minority over-sampling technique(SMOTE) is specifically designed for learning from imbalanced datasets, generating synthetic minority class examples by interpolating between minority class examples nearby. However, the SMOTE encounters the overgeneralization problem. The densitybased spatial clustering of applications with noise(DBSCAN) is not rigorous when dealing with the samples near the borderline.We optimize the DBSCAN algorithm for this problem to make clustering more reasonable. This paper integrates the optimized DBSCAN and SMOTE, and proposes a density-based synthetic minority over-sampling technique(DSMOTE). First, the optimized DBSCAN is used to divide the samples of the minority class into three groups, including core samples, borderline samples and noise samples, and then the noise samples of minority class is removed to synthesize more effective samples. In order to make full use of the information of core samples and borderline samples,different strategies are used to over-sample core samples and borderline samples. Experiments show that DSMOTE can achieve better results compared with SMOTE and Borderline-SMOTE in terms of precision, recall and F-value.展开更多
The conventional data envelopment analysis （DEA） measures the relative efficiencies of a set of decision making units with exact values of inputs and outputs. In real-world prob- lems, however, inputs and outputs ty...The conventional data envelopment analysis （DEA） measures the relative efficiencies of a set of decision making units with exact values of inputs and outputs. In real-world prob- lems, however, inputs and outputs typically have some levels of fuzziness. To analyze a decision making unit （DMU） with fuzzy input/output data, previous studies provided the fuzzy DEA model and proposed an associated evaluating approach. Nonetheless, numerous deficiencies must still be improved, including the α- cut approaches, types of fuzzy numbers, and ranking techniques. Moreover, a fuzzy sample DMU still cannot be evaluated for the Fuzzy DEA model. Therefore, this paper proposes a fuzzy DEA model based on sample decision making unit （FSDEA）. Five eval- uation approaches and the related algorithm and ranking methods are provided to test the fuzzy sample DMU of the FSDEA model. A numerical experiment is used to demonstrate and compare the results with those obtained using alternative approaches.展开更多
Data processing of small samples is an important and valuable research problem in the electronic equipment test. Because it is difficult and complex to determine the probability distribution of small samples, it is di...Data processing of small samples is an important and valuable research problem in the electronic equipment test. Because it is difficult and complex to determine the probability distribution of small samples, it is difficult to use the traditional probability theory to process the samples and assess the degree of uncertainty. Using the grey relational theory and the norm theory, the grey distance information approach, which is based on the grey distance information quantity of a sample and the average grey distance information quantity of the samples, is proposed in this article. The definitions of the grey distance information quantity of a sample and the average grey distance information quantity of the samples, with their characteristics and algorithms, are introduced. The correlative problems, including the algorithm of estimated value, the standard deviation, and the acceptance and rejection criteria of the samples and estimated results, are also proposed. Moreover, the information whitening ratio is introduced to select the weight algorithm and to compare the different samples. Several examples are given to demonstrate the application of the proposed approach. The examples show that the proposed approach, which has no demand for the probability distribution of small samples, is feasible and effective.展开更多
The lifting technique is now the most popular tool for dealing with sampled-data controlsystems. However, for the robust stability problem the system norm is not preserved by the liftingas expected. And the result is ...The lifting technique is now the most popular tool for dealing with sampled-data controlsystems. However, for the robust stability problem the system norm is not preserved by the liftingas expected. And the result is generally conservative under the small gain condition. The reason forthe norm di?erence by the lifting is that the state transition operator in the lifted system is zero inthis case. A new approach to the robust stability analysis is proposed. It is to use an equivalentdiscrete-time uncertainty to replace the continuous-time uncertainty. Then the general discretizedmethod can be used for the robust stability problem, and it is not conservative. Examples are givenin the paper.展开更多
The initial motivation of the lifting technique is to solve the H∞control problems. However, the conventional weighted H∞design does not meet the conditions required by lifting, so the result often leads to a misjud...The initial motivation of the lifting technique is to solve the H∞control problems. However, the conventional weighted H∞design does not meet the conditions required by lifting, so the result often leads to a misjudgement of the design. Two conditions required by using the lifting technique are presented based on the basic formulae of the lifting. It is pointed out that only the H∞disturbance attenuation problem with no weighting functions can meet these conditions, hence, the application of the lifting technique is quite limited.展开更多
This paper focuses on the fast rate fault detection filter (FDF) problem for a class of multirate sampled-data (MSD) systems. A lifting technique is used to convert such an MSD system into a linear time-invariant disc...This paper focuses on the fast rate fault detection filter (FDF) problem for a class of multirate sampled-data (MSD) systems. A lifting technique is used to convert such an MSD system into a linear time-invariant discrete-time one and an unknown input observer (UIO) is considered as FDF to generate residual. The design of FDF is formulated as an H∞ optimization problem and a solvable condition as well as an optimal solution are derived. The causality of the residual generator can be guaranteed so that the fast rate residual can be implemented via inverse lifting. A numerical example is included to demonstrate the feasibility of the obtained results.展开更多
A new approach is proposed for robust H2 problem of uncertain sampled-data systems. Through introducing a free variable, a new Lyapunov asymptotical stability criterion with less conservativeness is established. Based...A new approach is proposed for robust H2 problem of uncertain sampled-data systems. Through introducing a free variable, a new Lyapunov asymptotical stability criterion with less conservativeness is established. Based on this criterion, some sufficient conditions on two classes of robust H2 problems for uncertain sampled-data control systems axe presented through a set of coupled linear matrix inequalities. Finally, the less conservatism and potential of the developed results are illustrated via a numerical example.展开更多
This article investigates the problem of robust H∞ controller design for sampled-data systems with time-varying norm-bounded parameter uncertainties in the state matrices. Attention is focused on the design of a caus...This article investigates the problem of robust H∞ controller design for sampled-data systems with time-varying norm-bounded parameter uncertainties in the state matrices. Attention is focused on the design of a causal sampled-data controller, which guarantees the asymptotical stability of the closed-loop system and reduces the effect of the disturbance input on the controlled output to a prescribed H∞ performance bound for all admissible uncertainties. Sufficient condition for the solvability of the problem is established in terms of linear matrix inequalities （LMIs）. It is shown that the desired H∞ controller can be constructed by solving certain LMIs. An illustrative example is given to demonstrate the effectiveness of the proposed method.展开更多
处于改建阶段的智能变电站采样模式复杂,继电保护装置难以发现采样回路轻微异常,导致回路隐患暴露时间严重滞后。针对上述问题,分析改建时期智能变电站的采样模式和二次设备配置情况,提出基于同源录波数据比对的继电保护采样回路异常检...处于改建阶段的智能变电站采样模式复杂,继电保护装置难以发现采样回路轻微异常,导致回路隐患暴露时间严重滞后。针对上述问题,分析改建时期智能变电站的采样模式和二次设备配置情况,提出基于同源录波数据比对的继电保护采样回路异常检测方法。首先,利用双向编码器表征(bidirectional encoder representations from transformers,BERT)语言模型与余弦相似度算法,实现同源录波数据的通道匹配。然后,利用重采样技术和曼哈顿距离完成波形的采样频率统一与时域对齐。最后,基于动态时间规整(dynamic time warping,DTW)算法提出改进算法,并结合采样点偏移量共同设置采样回路的异常判据。算例分析表明,该方法可以完成录波数据的同源通道匹配,实现波形的一致性对齐,并且相比于传统DTW算法,改进DTW算法对异常状态识别的灵敏性和准确性更高。根据异常判据能够有效检测继电保护采样回路的异常状态,确保了智能变电站的安全可靠运行。展开更多
针对利用海量数据构建分类模型时训练数据规模大、训练时间长且碳排放量大的问题,提出面向低能耗高性能的分类器两阶段数据选择方法TSDS(Two-Stage Data Selection)。首先,通过修正余弦相似度确定聚类中心,并将样本数据进行基于不相似...针对利用海量数据构建分类模型时训练数据规模大、训练时间长且碳排放量大的问题,提出面向低能耗高性能的分类器两阶段数据选择方法TSDS(Two-Stage Data Selection)。首先,通过修正余弦相似度确定聚类中心,并将样本数据进行基于不相似点的分裂层次聚类;其次,对聚类结果按数据分布自适应抽样以组成高质量的子样本集;最后,利用子样本集在分类模型上训练,在加速训练过程的同时提升模型精度。在Spambase、Bupa和Phoneme等6个数据集上构建支持向量机(SVM)和多层感知机(MLP)分类模型,验证TSDS的性能。实验结果表明在样本数据压缩比达到85.00%的情况下,TSDS能将分类模型准确率提升3~10个百分点,同时加速模型训练,使训练SVM分类器的能耗平均降低93.76%,训练MLP分类器的能耗平均降低75.41%。可见,TSDS在大数据场景的分类任务上既能缩短训练时间和减少能耗,又能提升分类器性能,从而助力实现“双碳”目标。展开更多
基金supported by the National Key Research and Development Program of China(2018YFB1003700)the Scientific and Technological Support Project(Society)of Jiangsu Province(BE2016776)+2 种基金the“333” project of Jiangsu Province(BRA2017228 BRA2017401)the Talent Project in Six Fields of Jiangsu Province(2015-JNHB-012)
文摘For imbalanced datasets, the focus of classification is to identify samples of the minority class. The performance of current data mining algorithms is not good enough for processing imbalanced datasets. The synthetic minority over-sampling technique(SMOTE) is specifically designed for learning from imbalanced datasets, generating synthetic minority class examples by interpolating between minority class examples nearby. However, the SMOTE encounters the overgeneralization problem. The densitybased spatial clustering of applications with noise(DBSCAN) is not rigorous when dealing with the samples near the borderline.We optimize the DBSCAN algorithm for this problem to make clustering more reasonable. This paper integrates the optimized DBSCAN and SMOTE, and proposes a density-based synthetic minority over-sampling technique(DSMOTE). First, the optimized DBSCAN is used to divide the samples of the minority class into three groups, including core samples, borderline samples and noise samples, and then the noise samples of minority class is removed to synthesize more effective samples. In order to make full use of the information of core samples and borderline samples,different strategies are used to over-sample core samples and borderline samples. Experiments show that DSMOTE can achieve better results compared with SMOTE and Borderline-SMOTE in terms of precision, recall and F-value.
基金supported by the National Natural Science Foundation of China (70961005)211 Project for Postgraduate Student Program of Inner Mongolia University+1 种基金National Natural Science Foundation of Inner Mongolia (2010Zd342011MS1002)
文摘The conventional data envelopment analysis （DEA） measures the relative efficiencies of a set of decision making units with exact values of inputs and outputs. In real-world prob- lems, however, inputs and outputs typically have some levels of fuzziness. To analyze a decision making unit （DMU） with fuzzy input/output data, previous studies provided the fuzzy DEA model and proposed an associated evaluating approach. Nonetheless, numerous deficiencies must still be improved, including the α- cut approaches, types of fuzzy numbers, and ranking techniques. Moreover, a fuzzy sample DMU still cannot be evaluated for the Fuzzy DEA model. Therefore, this paper proposes a fuzzy DEA model based on sample decision making unit （FSDEA）. Five eval- uation approaches and the related algorithm and ranking methods are provided to test the fuzzy sample DMU of the FSDEA model. A numerical experiment is used to demonstrate and compare the results with those obtained using alternative approaches.
文摘Data processing of small samples is an important and valuable research problem in the electronic equipment test. Because it is difficult and complex to determine the probability distribution of small samples, it is difficult to use the traditional probability theory to process the samples and assess the degree of uncertainty. Using the grey relational theory and the norm theory, the grey distance information approach, which is based on the grey distance information quantity of a sample and the average grey distance information quantity of the samples, is proposed in this article. The definitions of the grey distance information quantity of a sample and the average grey distance information quantity of the samples, with their characteristics and algorithms, are introduced. The correlative problems, including the algorithm of estimated value, the standard deviation, and the acceptance and rejection criteria of the samples and estimated results, are also proposed. Moreover, the information whitening ratio is introduced to select the weight algorithm and to compare the different samples. Several examples are given to demonstrate the application of the proposed approach. The examples show that the proposed approach, which has no demand for the probability distribution of small samples, is feasible and effective.
文摘The lifting technique is now the most popular tool for dealing with sampled-data controlsystems. However, for the robust stability problem the system norm is not preserved by the liftingas expected. And the result is generally conservative under the small gain condition. The reason forthe norm di?erence by the lifting is that the state transition operator in the lifted system is zero inthis case. A new approach to the robust stability analysis is proposed. It is to use an equivalentdiscrete-time uncertainty to replace the continuous-time uncertainty. Then the general discretizedmethod can be used for the robust stability problem, and it is not conservative. Examples are givenin the paper.
基金Supported by the Harbin Engineering University Fund for Basic Projects (heuft06041)
文摘The initial motivation of the lifting technique is to solve the H∞control problems. However, the conventional weighted H∞design does not meet the conditions required by lifting, so the result often leads to a misjudgement of the design. Two conditions required by using the lifting technique are presented based on the basic formulae of the lifting. It is pointed out that only the H∞disturbance attenuation problem with no weighting functions can meet these conditions, hence, the application of the lifting technique is quite limited.
基金Supported by National Natural Science Foundation of P. R. China (60374021)the Natural Science Foundation of Shandong Province (Y2002G05)the Youth Scientists Foundation of Shandong Province (03BS091, 05BS01007) and Education Ministry Foundation of P. R. China (20050422036)
文摘This paper focuses on the fast rate fault detection filter (FDF) problem for a class of multirate sampled-data (MSD) systems. A lifting technique is used to convert such an MSD system into a linear time-invariant discrete-time one and an unknown input observer (UIO) is considered as FDF to generate residual. The design of FDF is formulated as an H∞ optimization problem and a solvable condition as well as an optimal solution are derived. The causality of the residual generator can be guaranteed so that the fast rate residual can be implemented via inverse lifting. A numerical example is included to demonstrate the feasibility of the obtained results.
基金Supported by State Key Program of National Natural Science Foundation of China （60934009） National Natural Science Foundations of China （60801048 60974062）
基金Supported by National Basic Research Program of China （973 Program） （2009CB320601）, National Natural Science Foundation of China （60774048, 60821063）, the Program for Cheung Kong Scholars, and the Research Fund for the Doctoral Program of China Higher Education （20070145015）
文摘这份报纸学习样品数据的问题为有变化时间的延期的不明确的连续时间的模糊大规模系统的可靠 H 夸张控制。第一，模糊夸张模型( FHM )被用来为某些复杂大规模系统建立模型，然后根据 Lyapunov 指导方法和大规模系统的分散的控制理论，线性 matrixine 质量( LMI )基于条件 arederived toguarantee H 性能不仅当所有控制部件正在操作很好时，而且面对一些可能的致动器失败。而且，致动器的精确失败参数没被要求，并且要求仅仅是失败参数的更低、上面的界限。条件依赖于时间延期的上面的界限，并且不依赖于变化时间的延期的衍生物。因此，获得的结果是不太保守的。最后，二个例子被提供说明设计过程和它的有效性。
文摘A new approach is proposed for robust H2 problem of uncertain sampled-data systems. Through introducing a free variable, a new Lyapunov asymptotical stability criterion with less conservativeness is established. Based on this criterion, some sufficient conditions on two classes of robust H2 problems for uncertain sampled-data control systems axe presented through a set of coupled linear matrix inequalities. Finally, the less conservatism and potential of the developed results are illustrated via a numerical example.
基金supported by the National Natural Science Foundation of China (60574004 60736024+1 种基金 60674043) the Key Project of Science and Technology Research of the Ministry of Education of China (708069).
文摘This article investigates the problem of robust H∞ controller design for sampled-data systems with time-varying norm-bounded parameter uncertainties in the state matrices. Attention is focused on the design of a causal sampled-data controller, which guarantees the asymptotical stability of the closed-loop system and reduces the effect of the disturbance input on the controlled output to a prescribed H∞ performance bound for all admissible uncertainties. Sufficient condition for the solvability of the problem is established in terms of linear matrix inequalities （LMIs）. It is shown that the desired H∞ controller can be constructed by solving certain LMIs. An illustrative example is given to demonstrate the effectiveness of the proposed method.
文摘处于改建阶段的智能变电站采样模式复杂,继电保护装置难以发现采样回路轻微异常,导致回路隐患暴露时间严重滞后。针对上述问题,分析改建时期智能变电站的采样模式和二次设备配置情况,提出基于同源录波数据比对的继电保护采样回路异常检测方法。首先,利用双向编码器表征(bidirectional encoder representations from transformers,BERT)语言模型与余弦相似度算法,实现同源录波数据的通道匹配。然后,利用重采样技术和曼哈顿距离完成波形的采样频率统一与时域对齐。最后,基于动态时间规整(dynamic time warping,DTW)算法提出改进算法,并结合采样点偏移量共同设置采样回路的异常判据。算例分析表明,该方法可以完成录波数据的同源通道匹配,实现波形的一致性对齐,并且相比于传统DTW算法,改进DTW算法对异常状态识别的灵敏性和准确性更高。根据异常判据能够有效检测继电保护采样回路的异常状态,确保了智能变电站的安全可靠运行。
文摘针对利用海量数据构建分类模型时训练数据规模大、训练时间长且碳排放量大的问题,提出面向低能耗高性能的分类器两阶段数据选择方法TSDS(Two-Stage Data Selection)。首先,通过修正余弦相似度确定聚类中心,并将样本数据进行基于不相似点的分裂层次聚类;其次,对聚类结果按数据分布自适应抽样以组成高质量的子样本集;最后,利用子样本集在分类模型上训练,在加速训练过程的同时提升模型精度。在Spambase、Bupa和Phoneme等6个数据集上构建支持向量机(SVM)和多层感知机(MLP)分类模型,验证TSDS的性能。实验结果表明在样本数据压缩比达到85.00%的情况下,TSDS能将分类模型准确率提升3~10个百分点,同时加速模型训练,使训练SVM分类器的能耗平均降低93.76%,训练MLP分类器的能耗平均降低75.41%。可见,TSDS在大数据场景的分类任务上既能缩短训练时间和减少能耗,又能提升分类器性能,从而助力实现“双碳”目标。
