三网融合下广电文本信息的分类技术研究

发布时间：2018-05-23 22:32

本文选题：三网融合 + 内容监管　；参考：《华中科技大学》2013年硕士论文

【摘要】：三网融合已逐渐成为我们信息技术发展的大趋势。在三网融合的大背景下，信息的传播途径将更加丰富、传播时间将更短、覆盖面将更大。由此会带来一系列的问题：大量的色情、暴力、血腥、反动、迷信的信息会得到更广泛、快速的传播，产生更大的危害；知识产权难以得到保护；垃圾信息充斥网络。因此需要一个强有力的内容监管系统对整个网络进行监管，屏蔽不良信息，过滤垃圾信息，打击盗版，净化网络环境，为全网内容可管、可控提供保障。随着网络文字媒体的不断涌现翻新，在三网融合的大前提下，文本信息监管愈发显的重要。文字内容监管系统是内容监管系统下的一个子系统，承担着对网络中一切文字信息的监管责任。文字监管系统通过对网络中的文字信息内容进行分析，对文本进行自动分类，自动识别，自动处理。面对海量的文字信息，要高效的进行处理，需要进行文本分类。运用文本分类方法能够屏蔽不良信息、过滤垃圾信息、保护知识产权。文本分类的效率会直接影响到文字内容监管系统的工作效率，因此，改进文本分类的算法是研究的重点。特征权重计算和分类模型是影响文本分类的关键因素。改进特征权重算法，使得更能表征文本类别的特征项获得更高的权重值；改进分类模型，使得分类的时间开销大幅降低。通过实验验证，改进后的方法准确度更高，速度更快，更加适合文字内容监管系统。
[Abstract]:The integration of three networks has gradually become the major trend of our information technology development. Under the background of three-network convergence, the information dissemination way will be richer, the dissemination time will be shorter, the coverage will be bigger. This will lead to a series of problems: a lot of pornography, violence, bloody, reactionary, superstitious information will be more widespread, rapid dissemination, causing greater harm; intellectual property rights are difficult to protect; spam information flooded the Internet. Therefore, we need a strong content supervision system to supervise the whole network, screen bad information, filter garbage information, crack down on piracy, purify the network environment, and provide security for the whole network content can be managed and controllable. With the continuous emergence and renovation of network text media, text information supervision becomes more and more important under the premise of the integration of three networks. The text content supervision system is a subsystem under the content supervision system, which is responsible for the supervision of all the text information in the network. By analyzing the content of the text information in the network, the text supervision system carries on the automatic classification, the automatic recognition, the automatic processing to the text. In the face of massive text information, text classification is needed to deal with it efficiently. Text classification method can screen bad information, filter garbage information and protect intellectual property rights. The efficiency of text classification will directly affect the efficiency of text content monitoring system, so improving the algorithm of text classification is the focus of research. Feature weight calculation and classification model are the key factors affecting text classification. By improving the feature weight algorithm, the feature items of text classification can be represented more efficiently, and the classification model can be improved to reduce the time cost of classification. Experimental results show that the improved method is more accurate, faster and more suitable for text content monitoring system.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.1

【参考文献】