针对目前网络异常流量识别准确率不高、基础深度学习模型特征提取能力不足以及循环神经网络训练效率低等问题,提出了基于卷积神经网络(Convolutional Neural Network,CNN)-双向内置注意力简单循环单元(Bidirectional Built in Attention...针对目前网络异常流量识别准确率不高、基础深度学习模型特征提取能力不足以及循环神经网络训练效率低等问题,提出了基于卷积神经网络(Convolutional Neural Network,CNN)-双向内置注意力简单循环单元(Bidirectional Built in Attention Simple Recurrent Unit,BiBASRU)-AT的网络异常流量检测模型。采用深层一维卷积模块提取流量局部特征表示,对高维度流量特征进行降维且学习到显著分类特征,增强模型的特征表示能力;同时构建内置自注意力简单循环单元(Built in self Attention Simple Recurrent Unit,BASRU)以同时捕捉流量中长距离的时序特征信息和内部特征之间的相互依赖关系,进一步挖掘流量特征内的高维结构信息。软注意力机制识别出对分类结果影响较大的重点特征,赋予关键特征更高权重,避免无关信息对分类结果造成干扰,最后由线性层输出分类概率分布,经Softmax函数归一化后取最大值对应标签作为流量识别结果。在多分类网络异常流量公开标准数据集UNSW-NB15上的实验结果表明,该模型取得了92.81%的F1值,高于实验对比的其他先进深度学习模型的结果,内置自注意力简单循环单元特征捕捉能力和训练效率优于其他传统循环神经网络的结果,证明了模型的可行性和有效性。展开更多
China's transition from a planned economy to a market economy has resulted in many changes in its urban structure and society and provided an opportunity for a quasi-longitudinal case study on the relationship bet...China's transition from a planned economy to a market economy has resulted in many changes in its urban structure and society and provided an opportunity for a quasi-longitudinal case study on the relationship between the built environment and activity-travel behavior.This paper draws upon data from an activity diary survey conducted in Beijing in 2007.The survey sample comprised 652 residents living in Danwei(work unit),commodity housing,and affordable housing neighborhoods.On the basis of the three-dimensional geo-visualization analysis of the space-time path and statistical multivariate regression models of daily travel and leisure time,it was found that both residential spatial factors and socio-demographics influence residents' daily behaviors.The findings show that Danwei residents have less daily travel time than those who live in commodity housing,but people living in affordable housing endure the longest travel time.Daily leisure time is associated more with individual attributes.We argue that although China's transition is currently gradual,the Danwei system may continue to play significant roles in daily life,and it might provide a valuable model for neighborhood spatial planning.展开更多
目的探讨OB内置式棉条联合穴位按摩预防危重症监护室(intensive care unit,ICU)大便失禁患者失禁相关性皮炎(incontinence-associated dermatitis,IAD)的效果。方法选择本院ICU收治的发生大便失禁的患者100例,采用随机数字表法将患者分...目的探讨OB内置式棉条联合穴位按摩预防危重症监护室(intensive care unit,ICU)大便失禁患者失禁相关性皮炎(incontinence-associated dermatitis,IAD)的效果。方法选择本院ICU收治的发生大便失禁的患者100例,采用随机数字表法将患者分为对照组和观察组,每组各50例。对照组患者采用常规护理方法,观察组在对照组基础上应用OB内置式棉条留置肛门联合穴位按摩。观察两组患者IAD发生率、发生时间、IAD严重程度及护理工作量。结果观察组患者IAD发生率低于对照组,IAD发生时间晚于对照组,IAD严重程度轻于对照组,患者失禁护理时间短于对照组,每天会阴清洗次数少于对照组,两组比较,均P<0.05,差异有统计学意义。结论对ICU大便失禁患者应用OB内置式棉条留置肛门联合穴位按摩,可降低患者IAD发生率及严重程度,减轻护理工作量,值得临床推广应用。展开更多
文摘针对目前网络异常流量识别准确率不高、基础深度学习模型特征提取能力不足以及循环神经网络训练效率低等问题,提出了基于卷积神经网络(Convolutional Neural Network,CNN)-双向内置注意力简单循环单元(Bidirectional Built in Attention Simple Recurrent Unit,BiBASRU)-AT的网络异常流量检测模型。采用深层一维卷积模块提取流量局部特征表示,对高维度流量特征进行降维且学习到显著分类特征,增强模型的特征表示能力;同时构建内置自注意力简单循环单元(Built in self Attention Simple Recurrent Unit,BASRU)以同时捕捉流量中长距离的时序特征信息和内部特征之间的相互依赖关系,进一步挖掘流量特征内的高维结构信息。软注意力机制识别出对分类结果影响较大的重点特征,赋予关键特征更高权重,避免无关信息对分类结果造成干扰,最后由线性层输出分类概率分布,经Softmax函数归一化后取最大值对应标签作为流量识别结果。在多分类网络异常流量公开标准数据集UNSW-NB15上的实验结果表明,该模型取得了92.81%的F1值,高于实验对比的其他先进深度学习模型的结果,内置自注意力简单循环单元特征捕捉能力和训练效率优于其他传统循环神经网络的结果,证明了模型的可行性和有效性。
基金Under the auspices of National Natural Science Foundation of China(No.40671058,41071102)National'TwelfthFive-Year'Plan for Science and Technology Support(No.2012BAJ 05B04)
文摘China's transition from a planned economy to a market economy has resulted in many changes in its urban structure and society and provided an opportunity for a quasi-longitudinal case study on the relationship between the built environment and activity-travel behavior.This paper draws upon data from an activity diary survey conducted in Beijing in 2007.The survey sample comprised 652 residents living in Danwei(work unit),commodity housing,and affordable housing neighborhoods.On the basis of the three-dimensional geo-visualization analysis of the space-time path and statistical multivariate regression models of daily travel and leisure time,it was found that both residential spatial factors and socio-demographics influence residents' daily behaviors.The findings show that Danwei residents have less daily travel time than those who live in commodity housing,but people living in affordable housing endure the longest travel time.Daily leisure time is associated more with individual attributes.We argue that although China's transition is currently gradual,the Danwei system may continue to play significant roles in daily life,and it might provide a valuable model for neighborhood spatial planning.
文摘目的探讨OB内置式棉条联合穴位按摩预防危重症监护室(intensive care unit,ICU)大便失禁患者失禁相关性皮炎(incontinence-associated dermatitis,IAD)的效果。方法选择本院ICU收治的发生大便失禁的患者100例,采用随机数字表法将患者分为对照组和观察组,每组各50例。对照组患者采用常规护理方法,观察组在对照组基础上应用OB内置式棉条留置肛门联合穴位按摩。观察两组患者IAD发生率、发生时间、IAD严重程度及护理工作量。结果观察组患者IAD发生率低于对照组,IAD发生时间晚于对照组,IAD严重程度轻于对照组,患者失禁护理时间短于对照组,每天会阴清洗次数少于对照组,两组比较,均P<0.05,差异有统计学意义。结论对ICU大便失禁患者应用OB内置式棉条留置肛门联合穴位按摩,可降低患者IAD发生率及严重程度,减轻护理工作量,值得临床推广应用。