随着Internet上信息量的飞速增长,成千上万的网上文档需要分类以方便用户的浏览和获取。因此文档的自动分类工作已经越来越受到重视,一些相应的分类方法也应运而生。但其中很少有涉及到"层次化"的分类领域,且绝大多数方法仅...随着Internet上信息量的飞速增长,成千上万的网上文档需要分类以方便用户的浏览和获取。因此文档的自动分类工作已经越来越受到重视,一些相应的分类方法也应运而生。但其中很少有涉及到"层次化"的分类领域,且绝大多数方法仅仅返回单个分类结果。文中,我们提出了一种新的文档自动分类方法:MRHC(Multicategory ReturnedAlgorithmforHierarchicalClassification)。该方法着眼于层次化的分类技术,并在适当的情况下为文档返回多个分类结果。该方法中结合了特征削减和增量学习技术以便提高分类性能。最后,为了更加准确、客观的评价分类结果,提出了一种新的评估方法:LEP(Length of Error Path)。实验结果表明,提出的分类方法响应时间短,分类准确度高,具有较强的实用性。展开更多
当今WiFi技术迅速发展,WiFi设备的种类和数量随之急剧增长,但WiFi设备识别这方面的研究并不多,仅有的一些研究也过多的依靠beacon节点主动搜集无线信号的方式开展。本文基于非侵入式监测搜集的WiFi信号数据,提出了一种WiFi设备类型的识...当今WiFi技术迅速发展,WiFi设备的种类和数量随之急剧增长,但WiFi设备识别这方面的研究并不多,仅有的一些研究也过多的依靠beacon节点主动搜集无线信号的方式开展。本文基于非侵入式监测搜集的WiFi信号数据,提出了一种WiFi设备类型的识别机制。通过监测WiFi通信过程中的技术参数,如接收信号强度(Received Signal Strength,RSS)、MAC地址(Media Access Control Address)、通信时间戳等,我们分析提取各类WiFi设备的特征,构建特征向量,然后运用机器学习中相对成熟的分类算法实现对常见无线设备如手机、笔记本电脑和无线路由器的分类。实验结果表明,本文提出的设备类型识别机制使用不同分类算法进行层次化分类后,均可达到较好的效果。展开更多
This paper focuses on some key problems in web community discovery and link analysis.Based on the topic-oriented technology,the characteristics of a bipartite graph are studied.An Х bipartite core set is introduced t...This paper focuses on some key problems in web community discovery and link analysis.Based on the topic-oriented technology,the characteristics of a bipartite graph are studied.An Х bipartite core set is introduced to more clearly define extracting ways.By scanning the topic subgraph to construct Х bipartite graph and then prune the graph with i and j ,an Х bipartite core set,which is also the minimum element of a community,can be found.Finally,a hierarchical clustering algorithm is applied to many Х bipartite core sets and the dendrogram of the community inner construction is obtained.The correctness of the constructing and pruning method is proved and the algorithm is designed.The typical datasets in the experiment are prepared according to the way in HITS(hyperlink-induced topic search).Ten topics and four search engines are chosen and the returned results are integrated.The modularity,which is a measure of the strength of the community structure in the social network,is used to validate the efficiency of the proposed method.The experimental results show that the proposed algorithm is effective and efficient.展开更多
文摘随着Internet上信息量的飞速增长,成千上万的网上文档需要分类以方便用户的浏览和获取。因此文档的自动分类工作已经越来越受到重视,一些相应的分类方法也应运而生。但其中很少有涉及到"层次化"的分类领域,且绝大多数方法仅仅返回单个分类结果。文中,我们提出了一种新的文档自动分类方法:MRHC(Multicategory ReturnedAlgorithmforHierarchicalClassification)。该方法着眼于层次化的分类技术,并在适当的情况下为文档返回多个分类结果。该方法中结合了特征削减和增量学习技术以便提高分类性能。最后,为了更加准确、客观的评价分类结果,提出了一种新的评估方法:LEP(Length of Error Path)。实验结果表明,提出的分类方法响应时间短,分类准确度高,具有较强的实用性。
文摘当今WiFi技术迅速发展,WiFi设备的种类和数量随之急剧增长,但WiFi设备识别这方面的研究并不多,仅有的一些研究也过多的依靠beacon节点主动搜集无线信号的方式开展。本文基于非侵入式监测搜集的WiFi信号数据,提出了一种WiFi设备类型的识别机制。通过监测WiFi通信过程中的技术参数,如接收信号强度(Received Signal Strength,RSS)、MAC地址(Media Access Control Address)、通信时间戳等,我们分析提取各类WiFi设备的特征,构建特征向量,然后运用机器学习中相对成熟的分类算法实现对常见无线设备如手机、笔记本电脑和无线路由器的分类。实验结果表明,本文提出的设备类型识别机制使用不同分类算法进行层次化分类后,均可达到较好的效果。
基金The National Natural Science Foundation of China(No.60773216)the National High Technology Research and Development Program of China(863Program)(No.2006AA010109)+1 种基金the Natural Science Foundation of Renmin University of China(No.06XNB052)Free Exploration Project(985 Project of Renmin University of China)(No.21361231)
文摘This paper focuses on some key problems in web community discovery and link analysis.Based on the topic-oriented technology,the characteristics of a bipartite graph are studied.An Х bipartite core set is introduced to more clearly define extracting ways.By scanning the topic subgraph to construct Х bipartite graph and then prune the graph with i and j ,an Х bipartite core set,which is also the minimum element of a community,can be found.Finally,a hierarchical clustering algorithm is applied to many Х bipartite core sets and the dendrogram of the community inner construction is obtained.The correctness of the constructing and pruning method is proved and the algorithm is designed.The typical datasets in the experiment are prepared according to the way in HITS(hyperlink-induced topic search).Ten topics and four search engines are chosen and the returned results are integrated.The modularity,which is a measure of the strength of the community structure in the social network,is used to validate the efficiency of the proposed method.The experimental results show that the proposed algorithm is effective and efficient.