A multi-class method is proposed based on Error Correcting Output Codes algorithm in order to get better performance of attack recognition in Wireless Sensor Networks. Aiming to enhance the accuracy of attack detectio...A multi-class method is proposed based on Error Correcting Output Codes algorithm in order to get better performance of attack recognition in Wireless Sensor Networks. Aiming to enhance the accuracy of attack detection, the multi-class method is constructed with Hadamard matrix and two-class Support Vector Machines. In order to minimize the complexity of the algorithm, sparse coding method is applied in this paper. The comprehensive experimental results show that this modified multi-class method has better attack detection rate compared with other three coding algorithms, and its time efficiency is higher than Hadamard coding algorithm.展开更多
Fuel injectors are considered as an important component of combustion engines. Operational weakness can possibly lead to the complete machine malfunction, decreasing reliability and leading to loss of production. To o...Fuel injectors are considered as an important component of combustion engines. Operational weakness can possibly lead to the complete machine malfunction, decreasing reliability and leading to loss of production. To overcome these circumstances, various condition monitoring techniques can be applied. The application of acoustic signals is common in the field of fault diagnosis of rotating machinery. Advanced signal processing is utilized for the construction of features that are specialized in detecting fuel injector faults. A performance comparison between novelty detection algorithms in the form of one-class classifiers is presented. The one-class classifiers that were tested included One-Class Support Vector Machine (OCSVM) and One-Class Self Organizing Map (OCSOM). The acoustic signals of fuel injectors in different operational conditions were processed for feature extraction. Features from all the signals were used as input to the one-class classifiers. The one-class classifiers were trained only with healthy fuel injector conditions and compared with new experimental data which belonged to different operational conditions that were not included in the training set so as to contribute to generalization. The results present the effectiveness of one-class classifiers for detecting faults in fuel injectors.展开更多
A new method to evaluate the fitness of the Bayesian networks according to the observed data is provided. The main advantage of this criterion is that it is suitable for both the complete and incomplete cases while th...A new method to evaluate the fitness of the Bayesian networks according to the observed data is provided. The main advantage of this criterion is that it is suitable for both the complete and incomplete cases while the others not. Moreover it facilitates the computation greatly. In order to reduce the search space, the notation of equivalent class proposed by David Chickering is adopted. Instead of using the method directly, the novel criterion, variable ordering, and equivalent class are combined,moreover the proposed mthod avoids some problems caused by the previous one. Later, the genetic algorithm which allows global convergence, lack in the most of the methods searching for Bayesian network is applied to search for a good model in thisspace. To speed up the convergence, the genetic algorithm is combined with the greedy algorithm. Finally, the simulation shows the validity of the proposed approach.展开更多
Goal of this research is to detect possible relations between animal-related attitudes and verbal aggressiveness as well as types combining such parameters. The sample collected in 2015 contains two adult education cl...Goal of this research is to detect possible relations between animal-related attitudes and verbal aggressiveness as well as types combining such parameters. The sample collected in 2015 contains two adult education classes equivalent to secondary school level (class A = 23 inmates and B = 12 inmates, all male) at a correctional facility. Questionnaires were used. Network analysis software (Visone) and conventional statistics (SPSS) are used for calculating network variables (indegree, outdegree, katz, pageranketc) and implementing Spearman test and Principal Component Analysis. Inmates who have adopted an animal-friendly value system and are too coward to react against torture of animals, maintain a repressed emotion. If they do not intervene and provoke, then they are also not targeted by others. No reaction against torture is also connected with a deep-rooted aggressiveness. Concerning superficial aggressiveness, a profile, whose characterize is multiple verbal aggressiveness, can be attributed to repressed emotions. A type is torturing and indifferently restricts his aggressiveness, as he can satisfy his need of dominance by being aggressive towards animals. A type of inmate who loves animals and reacts against their torture, presents the most restricted and relatively smooth aggressiveness, as he discharges his repressed emotions to this reaction. Under condition of indifference, keeping pets is not evidence of loving but of a need of companionship. As for the deep-rooted aggressiveness (over-extroversion), it does not seem to be triggered by any repression.展开更多
Learning Bayesian network structure is one of the most important branches in Bayesian network. The most popular graphical representative of a Bayesian network structure is an essential graph. This paper shows a combin...Learning Bayesian network structure is one of the most important branches in Bayesian network. The most popular graphical representative of a Bayesian network structure is an essential graph. This paper shows a combined algorithm according to the three rules for finding the essential graph of a given directed acyclic graph. Moreover, the complexity and advantages of this combined algorithm over others are also discussed. The aim of this paper is to present the proof of the correctness of the combined algorithm.展开更多
针对超临界锅炉运行中氧化膜动态生长特性难以实时监测导致的机组运行效能劣化问题,本文提出一种高效融合深度学习与生长机理的氧化膜厚度预测模型(VMD-NPDCLO-BiLSTM-SAMME),利用变分模态分解(variational mode decomposition,VMD)将...针对超临界锅炉运行中氧化膜动态生长特性难以实时监测导致的机组运行效能劣化问题,本文提出一种高效融合深度学习与生长机理的氧化膜厚度预测模型(VMD-NPDCLO-BiLSTM-SAMME),利用变分模态分解(variational mode decomposition,VMD)将原始序列分解为若干相对平稳的子序列,采用混沌-莱维神经种群动态优化(neural population dynamics with chaotic-levy optimization,NPDCLO)算法,构建具有最优超参数配置的双向长短期记忆神经网络(bidirectional long short-term memory,BiLSTM)模型,并使用多类指数损失函数渐进添加模型(stagewise additive modeling using a multi-class exponential loss function,SAMME)将多个NPDCLO-BiLSTM弱分类器组合,构建VMD-NPDCLO-BiLSTM-SAMME强分类器模型对氧化膜生成机理模型中的高温过热器壁温与烟温进行预测,最终利用预测结果融合机理模型以实现氧化膜厚度的实时精确估计。仿真实验结果表明:本文提出的模型与现有的BiLSTM-SAMME模型相比,壁温的平均绝对误差与均方根误差分别降低32.52%、32.26%,烟温的平均绝对误差与均方根误差分别降低47.38%、55.27%;氧化膜厚度预测模型的平均误差为7.42%,验证了模型的有效性及工程适用性。展开更多
文摘A multi-class method is proposed based on Error Correcting Output Codes algorithm in order to get better performance of attack recognition in Wireless Sensor Networks. Aiming to enhance the accuracy of attack detection, the multi-class method is constructed with Hadamard matrix and two-class Support Vector Machines. In order to minimize the complexity of the algorithm, sparse coding method is applied in this paper. The comprehensive experimental results show that this modified multi-class method has better attack detection rate compared with other three coding algorithms, and its time efficiency is higher than Hadamard coding algorithm.
文摘Fuel injectors are considered as an important component of combustion engines. Operational weakness can possibly lead to the complete machine malfunction, decreasing reliability and leading to loss of production. To overcome these circumstances, various condition monitoring techniques can be applied. The application of acoustic signals is common in the field of fault diagnosis of rotating machinery. Advanced signal processing is utilized for the construction of features that are specialized in detecting fuel injector faults. A performance comparison between novelty detection algorithms in the form of one-class classifiers is presented. The one-class classifiers that were tested included One-Class Support Vector Machine (OCSVM) and One-Class Self Organizing Map (OCSOM). The acoustic signals of fuel injectors in different operational conditions were processed for feature extraction. Features from all the signals were used as input to the one-class classifiers. The one-class classifiers were trained only with healthy fuel injector conditions and compared with new experimental data which belonged to different operational conditions that were not included in the training set so as to contribute to generalization. The results present the effectiveness of one-class classifiers for detecting faults in fuel injectors.
基金This project was supported by the National Natural Science Foundation of China (70572045).
文摘A new method to evaluate the fitness of the Bayesian networks according to the observed data is provided. The main advantage of this criterion is that it is suitable for both the complete and incomplete cases while the others not. Moreover it facilitates the computation greatly. In order to reduce the search space, the notation of equivalent class proposed by David Chickering is adopted. Instead of using the method directly, the novel criterion, variable ordering, and equivalent class are combined,moreover the proposed mthod avoids some problems caused by the previous one. Later, the genetic algorithm which allows global convergence, lack in the most of the methods searching for Bayesian network is applied to search for a good model in thisspace. To speed up the convergence, the genetic algorithm is combined with the greedy algorithm. Finally, the simulation shows the validity of the proposed approach.
文摘Goal of this research is to detect possible relations between animal-related attitudes and verbal aggressiveness as well as types combining such parameters. The sample collected in 2015 contains two adult education classes equivalent to secondary school level (class A = 23 inmates and B = 12 inmates, all male) at a correctional facility. Questionnaires were used. Network analysis software (Visone) and conventional statistics (SPSS) are used for calculating network variables (indegree, outdegree, katz, pageranketc) and implementing Spearman test and Principal Component Analysis. Inmates who have adopted an animal-friendly value system and are too coward to react against torture of animals, maintain a repressed emotion. If they do not intervene and provoke, then they are also not targeted by others. No reaction against torture is also connected with a deep-rooted aggressiveness. Concerning superficial aggressiveness, a profile, whose characterize is multiple verbal aggressiveness, can be attributed to repressed emotions. A type is torturing and indifferently restricts his aggressiveness, as he can satisfy his need of dominance by being aggressive towards animals. A type of inmate who loves animals and reacts against their torture, presents the most restricted and relatively smooth aggressiveness, as he discharges his repressed emotions to this reaction. Under condition of indifference, keeping pets is not evidence of loving but of a need of companionship. As for the deep-rooted aggressiveness (over-extroversion), it does not seem to be triggered by any repression.
基金Supported by the National Natural Science Foundation of China (No. 60974082)
文摘Learning Bayesian network structure is one of the most important branches in Bayesian network. The most popular graphical representative of a Bayesian network structure is an essential graph. This paper shows a combined algorithm according to the three rules for finding the essential graph of a given directed acyclic graph. Moreover, the complexity and advantages of this combined algorithm over others are also discussed. The aim of this paper is to present the proof of the correctness of the combined algorithm.
文摘针对超临界锅炉运行中氧化膜动态生长特性难以实时监测导致的机组运行效能劣化问题,本文提出一种高效融合深度学习与生长机理的氧化膜厚度预测模型(VMD-NPDCLO-BiLSTM-SAMME),利用变分模态分解(variational mode decomposition,VMD)将原始序列分解为若干相对平稳的子序列,采用混沌-莱维神经种群动态优化(neural population dynamics with chaotic-levy optimization,NPDCLO)算法,构建具有最优超参数配置的双向长短期记忆神经网络(bidirectional long short-term memory,BiLSTM)模型,并使用多类指数损失函数渐进添加模型(stagewise additive modeling using a multi-class exponential loss function,SAMME)将多个NPDCLO-BiLSTM弱分类器组合,构建VMD-NPDCLO-BiLSTM-SAMME强分类器模型对氧化膜生成机理模型中的高温过热器壁温与烟温进行预测,最终利用预测结果融合机理模型以实现氧化膜厚度的实时精确估计。仿真实验结果表明:本文提出的模型与现有的BiLSTM-SAMME模型相比,壁温的平均绝对误差与均方根误差分别降低32.52%、32.26%,烟温的平均绝对误差与均方根误差分别降低47.38%、55.27%;氧化膜厚度预测模型的平均误差为7.42%,验证了模型的有效性及工程适用性。
