Advanced Persistent Threats(APTs)represent one of the most complex and dangerous categories of cyber-attacks characterised by their stealthy behaviour,long-term persistence,and ability to bypass traditional detection ...Advanced Persistent Threats(APTs)represent one of the most complex and dangerous categories of cyber-attacks characterised by their stealthy behaviour,long-term persistence,and ability to bypass traditional detection systems.The complexity of real-world network data poses significant challenges in detection.Machine learning models have shown promise in detecting APTs;however,their performance often suffers when trained on large datasets with redundant or irrelevant features.This study presents a novel,hybrid feature selection method designed to improve APT detection by reducing dimensionality while preserving the informative characteristics of the data.It combines Mutual Information(MI),Symmetric Uncertainty(SU)and Minimum Redundancy Maximum Relevance(mRMR)to enhance feature selection.MI and SU assess feature relevance,while mRMR maximises relevance and minimises redundancy,ensuring that the most impactful features are prioritised.This method addresses redundancy among selected features,improving the overall efficiency and effectiveness of the detection model.Experiments on a real-world APT datasets were conducted to evaluate the proposed method.Multiple classifiers including,Random Forest,Support Vector Machine(SVM),Gradient Boosting,and Neural Networks were used to assess classification performance.The results demonstrate that the proposed feature selection method significantly enhances detection accuracy compared to baseline models trained on the full feature set.The Random Forest algorithm achieved the highest performance,with near-perfect accuracy,precision,recall,and F1 scores(99.97%).The proposed adaptive thresholding algorithm within the selection method allows each classifier to benefit from a reduced and optimised feature space,resulting in improved training and predictive performance.This research offers a scalable and classifier-agnostic solution for dimensionality reduction in cybersecurity applications.展开更多
The topology control strategies of wireless sensor networks are very important for reducing the energy consumption of sensor nodes and prolonging the life-span of networks. In this paper, we put forward a minimum-ener...The topology control strategies of wireless sensor networks are very important for reducing the energy consumption of sensor nodes and prolonging the life-span of networks. In this paper, we put forward a minimum-energy path-preserving topology control （MPTC） algorithm based on a concept of none k-redundant edges. MPTC not only resolves the problem of excessive energy consumption because of the unclosed region in small minimum-energy communication network （SMECN）, but also preserves at least one minimum-energy path between every pair of nodes in a wireless sensor network. We also propose an energy-efficient reconfiguration protocol that maintains the minimum-energy path property in the case where the network topology changes dynamically. Finally, we demonstrate the performance improvements of our algorithm through simulation.展开更多
In order to overcome the shortcomings of the previous obstacle avoidance algorithms,an obstacle avoidance algorithm applicable to multiple mobile obstacles was proposed.The minimum prediction distance between obstacle...In order to overcome the shortcomings of the previous obstacle avoidance algorithms,an obstacle avoidance algorithm applicable to multiple mobile obstacles was proposed.The minimum prediction distance between obstacles and a manipulator was obtained according to the states of obstacles and transformed to escape velocity of the corresponding link of the manipulator.The escape velocity was introduced to the gradient projection method to obtain the joint velocity of the manipulator so as to complete the obstacle avoidance trajectory planning.A7-DOF manipulator was used in the simulation,and the results verified the effectiveness of the algorithm.展开更多
随着大量分布式能源的接入,配电系统的运行与控制方式愈加复杂。针对配电网状态估计方法面临分布式电源波动数据辨识困难、估计精度低、鲁棒性与估计时效性差等问题,提出一种基于集成深度神经网络的配电网分布式状态估计方法。首先,利...随着大量分布式能源的接入,配电系统的运行与控制方式愈加复杂。针对配电网状态估计方法面临分布式电源波动数据辨识困难、估计精度低、鲁棒性与估计时效性差等问题,提出一种基于集成深度神经网络的配电网分布式状态估计方法。首先,利用量测数据相关性检验的数据辨识技术识别不良数据和新能源波动数据。在此基础上,利用时域卷积网络(temporal convolutional network,TCN)-双向长短期记忆网络(bidirectional long short term memory,BILSTM)对不良数据进行修正。然后,建立集成深度神经网络(deep neural network,DNN)状态估计模型,采用最大相关-最小冗余(maximum relevance-minimum redundancy,MRMR)的方法优化训练样本,从而提高状态估计的精度和鲁棒性。最后,建立分布式集成深度神经网络模型,弥补了集中式状态估计速度慢的不足,从而提高状态估计效率。基于IEEE123配电网的算例分析表明,所提方法能更准确地辨识分布式电源波动数据和不良数据,同时提高状态估计的精度和效率,且具有较高的鲁棒性。展开更多
文摘由于具有高时间分辨率、无创性,脑电(Electroencephalogram,EEG)信号被广泛应用于航空航天任务操作员的疲劳、脑力负荷分析等。针对EEG信号多通道且各通道内信息不完全相同的特性,提出了一种基于最小冗余最大相关性(Minimum redundancy maximum relevance,mRMR)算法的EEG特征评价技术。通过设置目标变量,计算各通道内EEG特征与目标变量的互信息量、特征在通道内部的冗余度,可对EEG特征的性能做出评价。进一步,获取管制员在不同脑力负荷下的EEG数据,对一系列EEG特征做出评价并与已有研究、特征在不同分类方式下的可分性进行对比,验证了该特征评价技术的有效性。与现有的技术相比,该技术避免了灰色关联分析法确定权重参数和灰色关联度的主观性、避免了分类器评价法的差异性。相较于已有的特征选择算法,考虑了通道内部信息的冗余,使得评价结果更为准确。相较于基于统计学的相关技术,该方法可对特征的性能做出定量的评价,以便对不同指标进行比较。最后,阐述了该评价方式疲劳程度分析、情绪识别等方面的应用。
基金funded by Universiti Teknologi Malaysia under the UTM RA ICONIC Grant(Q.J130000.4351.09G61).
文摘Advanced Persistent Threats(APTs)represent one of the most complex and dangerous categories of cyber-attacks characterised by their stealthy behaviour,long-term persistence,and ability to bypass traditional detection systems.The complexity of real-world network data poses significant challenges in detection.Machine learning models have shown promise in detecting APTs;however,their performance often suffers when trained on large datasets with redundant or irrelevant features.This study presents a novel,hybrid feature selection method designed to improve APT detection by reducing dimensionality while preserving the informative characteristics of the data.It combines Mutual Information(MI),Symmetric Uncertainty(SU)and Minimum Redundancy Maximum Relevance(mRMR)to enhance feature selection.MI and SU assess feature relevance,while mRMR maximises relevance and minimises redundancy,ensuring that the most impactful features are prioritised.This method addresses redundancy among selected features,improving the overall efficiency and effectiveness of the detection model.Experiments on a real-world APT datasets were conducted to evaluate the proposed method.Multiple classifiers including,Random Forest,Support Vector Machine(SVM),Gradient Boosting,and Neural Networks were used to assess classification performance.The results demonstrate that the proposed feature selection method significantly enhances detection accuracy compared to baseline models trained on the full feature set.The Random Forest algorithm achieved the highest performance,with near-perfect accuracy,precision,recall,and F1 scores(99.97%).The proposed adaptive thresholding algorithm within the selection method allows each classifier to benefit from a reduced and optimised feature space,resulting in improved training and predictive performance.This research offers a scalable and classifier-agnostic solution for dimensionality reduction in cybersecurity applications.
基金supported by by National Natural Science Founda-tion of China (No. 60702055)Program for New Century ExcellentTalents in University (NCET-07-0914)the Science and Technology Research Project of Chongqing Municipal Education Commission of China (KJ070521)
文摘The topology control strategies of wireless sensor networks are very important for reducing the energy consumption of sensor nodes and prolonging the life-span of networks. In this paper, we put forward a minimum-energy path-preserving topology control （MPTC） algorithm based on a concept of none k-redundant edges. MPTC not only resolves the problem of excessive energy consumption because of the unclosed region in small minimum-energy communication network （SMECN）, but also preserves at least one minimum-energy path between every pair of nodes in a wireless sensor network. We also propose an energy-efficient reconfiguration protocol that maintains the minimum-energy path property in the case where the network topology changes dynamically. Finally, we demonstrate the performance improvements of our algorithm through simulation.
基金Supported by Ministeral Level Advanced Research Foundation(65822576)Beijing Municipal Education Commission(KM201310858004,KM201310858001)
文摘In order to overcome the shortcomings of the previous obstacle avoidance algorithms,an obstacle avoidance algorithm applicable to multiple mobile obstacles was proposed.The minimum prediction distance between obstacles and a manipulator was obtained according to the states of obstacles and transformed to escape velocity of the corresponding link of the manipulator.The escape velocity was introduced to the gradient projection method to obtain the joint velocity of the manipulator so as to complete the obstacle avoidance trajectory planning.A7-DOF manipulator was used in the simulation,and the results verified the effectiveness of the algorithm.
文摘随着大量分布式能源的接入,配电系统的运行与控制方式愈加复杂。针对配电网状态估计方法面临分布式电源波动数据辨识困难、估计精度低、鲁棒性与估计时效性差等问题,提出一种基于集成深度神经网络的配电网分布式状态估计方法。首先,利用量测数据相关性检验的数据辨识技术识别不良数据和新能源波动数据。在此基础上,利用时域卷积网络(temporal convolutional network,TCN)-双向长短期记忆网络(bidirectional long short term memory,BILSTM)对不良数据进行修正。然后,建立集成深度神经网络(deep neural network,DNN)状态估计模型,采用最大相关-最小冗余(maximum relevance-minimum redundancy,MRMR)的方法优化训练样本,从而提高状态估计的精度和鲁棒性。最后,建立分布式集成深度神经网络模型,弥补了集中式状态估计速度慢的不足,从而提高状态估计效率。基于IEEE123配电网的算例分析表明,所提方法能更准确地辨识分布式电源波动数据和不良数据,同时提高状态估计的精度和效率,且具有较高的鲁棒性。
