The article holds that the problem of the relationship between"words"and"meaning"has always been the philosophical proposition.Through analyzing the relationship between"words"and"me...The article holds that the problem of the relationship between"words"and"meaning"has always been the philosophical proposition.Through analyzing the relationship between"words"and"meaning"and probing into"disoourse","text"and"space-time history",the author thinks that the interpreters'preconceptions,preexistence,pre-structure and different perspectives have resulted in diverse forms of interpretations,and the being of each kind of form has its own rationality.Moreover,the process in which the interpreters interpret"words","discourse","text"and"history"is also the process to interpret themselves.Nevertheless,being able to say and being unable to say are soaked in the whole process of interpretation,and,as a result,human beings will always be confronted with a kind of"say and cannot say"embarrassment.展开更多
为了对饮食文本信息高效分类,建立一种基于word2vec和长短期记忆网络(Long-short term memory,LSTM)的分类模型。针对食物百科和饮食健康文本特点,首先利用word2vec实现包含语义信息的词向量表示,并解决了传统方法导致数据表示稀疏及维...为了对饮食文本信息高效分类,建立一种基于word2vec和长短期记忆网络(Long-short term memory,LSTM)的分类模型。针对食物百科和饮食健康文本特点,首先利用word2vec实现包含语义信息的词向量表示,并解决了传统方法导致数据表示稀疏及维度灾难问题,基于K-means++根据语义关系聚类以提高训练数据质量。由word2vec构建文本向量作为LSTM的初始输入,训练LSTM分类模型,自动提取特征,进行饮食宜、忌的文本分类。实验采用48 000个文档进行测试,结果显示,分类准确率为98.08%,高于利用tf-idf、bag-of-words等文本数值化表示方法以及基于支持向量机(Support vector machine,SVM)和卷积神经网络(Convolutional neural network,CNN)分类算法结果。实验结果表明,利用该方法能够高质量地对饮食文本自动分类,帮助人们有效地利用健康饮食信息。展开更多
传统词袋(bag of words,BoW)模型在构造视觉词典时一般采用k-means聚类方法实现,但k-means聚类方法的性能在很大程度上依赖于初始点的选择,从而导致生成的视觉词典鲁棒性较差,此外,每次迭代都要计算数据点与中心点的距离,计算复杂度高...传统词袋(bag of words,BoW)模型在构造视觉词典时一般采用k-means聚类方法实现,但k-means聚类方法的性能在很大程度上依赖于初始点的选择,从而导致生成的视觉词典鲁棒性较差,此外,每次迭代都要计算数据点与中心点的距离,计算复杂度高。针对上述问题,提出了一种改进的k-means聚类视觉词典构造方法,该方法首先对初始值的选取进行了优化,克服了随机选取初始值对聚类性能的影响,其次基于三角形不等式对计算进行了简化,使生成的视觉词典更加稳定,计算复杂度更低,最后引入权值分布对图像进行基于视觉词典的表示,并将基于改进的视觉词典的词袋模型应用于图像分类,提高了分类性能。通过在Caltech 101和Caltech 256两个数据库进行实验,验证了本文方法的有效性,并分析了词典库大小对分类性能的影响。从实验结果可以看出,采用本文方法所得到的分类正确率提高了5%～8%。展开更多
针对移动自组网(MANET,mobile ad hoc networks)入侵检测过程中的攻击类型多样性和监测数据海量性问题,提出了一种基于改进k-means算法的MANET异常检测方法。通过引入划分贡献度的概念,可合理地计算各维特征在检测中占有的权重,并将遗...针对移动自组网(MANET,mobile ad hoc networks)入侵检测过程中的攻击类型多样性和监测数据海量性问题,提出了一种基于改进k-means算法的MANET异常检测方法。通过引入划分贡献度的概念,可合理地计算各维特征在检测中占有的权重,并将遗传算法与快速聚类检测算法k-means相结合,解决了聚类检测结果容易陷入局部最优的问题,进而,提出了以上检测算法在Map Reduce框架下的设计方案,利用种群迁移策略在分布式处理器上实现了并行聚类检测。实验结果证明了该方法的检测准确率和运行效率均优于传统聚类检测方法。展开更多
基金the research achievements of 2013 Classroom Teaching Reform Project in Higher Education fiscally aided by Zhejiang province(Grant No.kg2013416)2014 Pre-research for High-Level Project in Humanities and Social Sciences financially supported by Huzhou University(Grant No.2014SKYY07)2015 Monographic Research Project financed by Zhejiang Association of Foreign Languages&Literatures(Grant No.ZWYB2015003)
文摘The article holds that the problem of the relationship between"words"and"meaning"has always been the philosophical proposition.Through analyzing the relationship between"words"and"meaning"and probing into"disoourse","text"and"space-time history",the author thinks that the interpreters'preconceptions,preexistence,pre-structure and different perspectives have resulted in diverse forms of interpretations,and the being of each kind of form has its own rationality.Moreover,the process in which the interpreters interpret"words","discourse","text"and"history"is also the process to interpret themselves.Nevertheless,being able to say and being unable to say are soaked in the whole process of interpretation,and,as a result,human beings will always be confronted with a kind of"say and cannot say"embarrassment.
文摘传统词袋(bag of words,BoW)模型在构造视觉词典时一般采用k-means聚类方法实现,但k-means聚类方法的性能在很大程度上依赖于初始点的选择,从而导致生成的视觉词典鲁棒性较差,此外,每次迭代都要计算数据点与中心点的距离,计算复杂度高。针对上述问题,提出了一种改进的k-means聚类视觉词典构造方法,该方法首先对初始值的选取进行了优化,克服了随机选取初始值对聚类性能的影响,其次基于三角形不等式对计算进行了简化,使生成的视觉词典更加稳定,计算复杂度更低,最后引入权值分布对图像进行基于视觉词典的表示,并将基于改进的视觉词典的词袋模型应用于图像分类,提高了分类性能。通过在Caltech 101和Caltech 256两个数据库进行实验,验证了本文方法的有效性,并分析了词典库大小对分类性能的影响。从实验结果可以看出,采用本文方法所得到的分类正确率提高了5%～8%。
文摘针对移动自组网(MANET,mobile ad hoc networks)入侵检测过程中的攻击类型多样性和监测数据海量性问题,提出了一种基于改进k-means算法的MANET异常检测方法。通过引入划分贡献度的概念,可合理地计算各维特征在检测中占有的权重,并将遗传算法与快速聚类检测算法k-means相结合,解决了聚类检测结果容易陷入局部最优的问题,进而,提出了以上检测算法在Map Reduce框架下的设计方案,利用种群迁移策略在分布式处理器上实现了并行聚类检测。实验结果证明了该方法的检测准确率和运行效率均优于传统聚类检测方法。
基金Supported by National Natural Science Foundation of China(10571036)the Key Discipline Development Program of Beijing Municipal Commission (XK100080537)
