面向增量同生主题的维吾尔文爬虫的研究  

Research for Uyghur crawler face to incremental contemporaneous topic

作　　者：赵永霄 哈力旦.阿布都热依木 张振东[1] 

机构地区：[1]新疆大学电气工程学院,乌鲁木齐830047

出　　处：《计算机应用研究》2014年第11期3269-3272,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(61163026;60865001)

摘　　要：针对传统的主题爬虫对网页信息缺乏在知识层面上的处理和理解的问题进行了研究,提出了一种面向增量同生主题的维吾尔文爬虫,通过建立一个增量主题词库优化传统的主题模型,来描述维吾尔文关键词的应用语境及场景,提高了计算网页相关度的准确率。用改进的IC主题敏感算法来预测子页面优先级,过滤无关的网页地址。依据上述方法编写爬虫系统,用构建的维吾尔文语料库进行实验,表明了基于此模型的爬虫具有更好的稳定性和准确度。According to the traditional topic crawler lacks of processing and understanding to Web information in the knowledge level,this paper proposed a Uyghur crawler face to incremental contemporaneous topic.It set up an incremental topic words li-brary to improve the topic model,to describe the application scenarios and context of the keywords.It improved the accuracy of Web relevancy.It calculated the priority of the page with the improved sensitive IC_PageRank algorithm and filtered the irrele-vant Web URL.On the basis of the above methods it made the crawler system.It did experiments based on the building of Uyghur corpus,the results show that the crawler based on this model has better stability and accuracy.

关 键 词：网络爬虫 网页分类 IC主题模型 锚文本 维吾尔文 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]