社会网络中不良URL的研究

发布时间：2018-03-22 13:45

本文选题：社会网络　切入点：不良URL　出处：《山东大学》2014年硕士论文　论文类型：学位论文

【摘要】：近年来,问答类社会网络迅速发展,用户量快速增长。经过数年的积累,以人为核心的问答类社会网络正成为互联网的主体应用之一。然而,随着社会网络的蓬勃发展,互联网安全受到了新的挑战。首先是用户信息的安全。在社会网络中,为了更好的交流,用户一般都会公布自己的个人信息。然而由于缺乏安全防护意识和未采取隐私措施,用户信息资料经常被非法公开或被一些不法分子非法利用。使得用户的隐私和安全问题频频发生。其次是不良URL的传播。由于社会网络中信息的快速传播性,很多黑客可以用社会网络传播不良信息。其中危害性比较大的是不良URL,主要有恶意URL、钓鱼URL。垃圾广告URL、色情URL等。这些给用户使用社会网络及社会网络的健康发展造成了严重的危害。本文是当前研究问答类社会网络中的不良URL的少数文章之一。有效的弥补了该类研究的不足。针对当前问答类社会网络中出现的不良URL现象进行了深入的研究。本文首先对社会网络中的不良URL的问题进行了分析。说出了其中的危害：传播恶意URL(包括钓鱼URL)、给用户造成威胁、安装恶意软件、窃听用户信息、盗取用户密码等威胁；或者传播色情URL,影响网络生态环境；或者散步大量广告,使用户难以找到自己需要的答案,或者用户收到许多不良信息的误导和骚扰。本文针对问答类社会网络中的问题,提出了解决方案。即首先爬取问答类社会网站,提取其中用户回答问题所发布的URL,然后通过urlvoid网站进行判断,可以迅速的找出该URL的性质,进而判定是否是恶意的URL。对于广告类的URL,也就是与用户问题不想对应的URL。通过文本相似度来计算URL与问题的匹配值。其中用到了网页关键字提取,文本相似度计算等。如果匹配值较低,则认为他们是不相关的,进而判断为不相关URL。最后实现实验并对实验结果进行评估。通过大量的实验及评估,实验取得了较好的实验结果。本文主要通过对问答类社会网络中不良URL的研究,通过对雅虎问答的爬取,分析,得到了以下结论,取得了一下成果： 1,本文是当前对社会网络中的不良URL研究的少数文章之一,针对问答类社会网络中出现的安全问题,提出了解决框架,并设计了我们的系统,通过了实验验证； 2,用雅虎S4平台提高了鉴定不良URL的速率,获得了较为理想的速率； 3,利用文本相似度处理,鉴定垃圾URL,广告URL,扩展了文本相似度原理的应用； 4,能够快速的判别出恶意URL,这大大减少了恶意URL的危害。
[Abstract]:In recent years, question-and-answer social network has developed rapidly and the number of users has increased rapidly. After several years of accumulation, the question-and-answer social network with people as the core is becoming one of the main applications of the Internet. However, with the rapid development of social network, Internet security is facing new challenges. First, the security of user information. In social networks, in order to communicate better, users generally publish their personal information. However, due to lack of security awareness and privacy measures, User information materials are often illegally exposed or illegally exploited by some lawless elements, which make the privacy and security problems of users occur frequently. Secondly, the spread of bad URL. Because of the rapid dissemination of information in social networks, Many hackers can use social networks to spread bad information. Among them, bad URLs are more harmful, mainly malicious URLs, phishing URL. spam URL. pornographic URL. These give users the use of social networks and social network health hair. The exhibition caused serious harm. This paper is one of the few articles of the current research on the bad URL in the social network of question and answer. It effectively makes up for the deficiency of this kind of research. This paper makes a deep research on the phenomenon of bad URL in the social network of question and answer at present. In this paper, we first analyze the problem of bad URL in social network, and point out the harm: spreading malicious URLs (including phishing URLs) is a threat to users. Install malicious software, eavesdrop on user information, steal user passwords and other threats; or spread pornography to affect the ecological environment of the network; or walk through a large number of advertisements that make it difficult for users to find the answers they need. Or the user receives a lot of bad information to mislead and harass. This paper puts forward a solution to the problems in the question answering social network, that is, first crawling the question answering social website, extracting the URLs published by the user answering the questions, and then judging by the urlvoid website. Can quickly find out the nature of the URL, and then determine whether the malicious URL. For the advertising class URL, that is, the user does not want to corresponding URL. through the text similarity to calculate the matching value between the URL and the problem. If the matching value is low, they are considered irrelevant and judged as irrelevant URLLs. Finally, the experiment is carried out and the experimental results are evaluated. Through a large number of experiments and evaluations, good experimental results are obtained. This article mainly through the question and answer social network bad URL research, through the Yahoo question and answer crawling, the analysis, has obtained the following conclusion, has obtained the achievement:. 1. This paper is one of the few articles on the research of the bad URL in social network. Aiming at the security problems in the social network of question and answer, the framework is proposed, and our system is designed and verified by experiment. 2, using Yahoo S4 platform to improve the rate of identification of bad URL, obtained a more ideal rate; 3. The application of text similarity principle is extended by using text similarity processing, identifying spam URLand advertising URLL; 4, can quickly identify malicious URL, which greatly reduces the harm of malicious URL.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09;TP309

【参考文献】