用户建模和事件检测技术在论坛监测系统中的研究与应用

发布时间：2018-07-12 07:53

本文选题：兴趣模型 + 用户关系　；参考：《浙江大学》2012年硕士论文

【摘要】：信息时代的飞速发展使得越来越多的人喜欢在互联网上交流和分享信息,网络论坛作为互联网用户交流的平台,从某种程度上改变了舆论的形成模式、传播方式和扩散速度,同时也丰富了人们的精神生活。但是,由于网络传播的自由性、交互性和开放性,使得网络论坛成为敏感信息以及广告泛滥的高发地带,如何把握网络论坛舆论发展的方向,塑造一个干净积极的互联网沟通平台,成为亟待解决的问题。本文就网络论坛中用户兴趣模型建立以及突发事件发现两方面展开深入研究,旨在帮助论坛管理员管理论坛中的用户和主题,促进论坛的良性发展。用户兴趣模型的建立是为了发现、提取和表达用户兴趣,通过合理的展现用户兴趣,使管理员更好的理解用户的关注点并做出相对应的处理。本文提出一种基于用户兴趣模型的用户监测方法,通过分析用户对敏感主题的关注度和用户关系模型,综合地定义用户的敏感度,从而判断用户的潜在行为。突发事件是指在短时间内被广泛传播的内容。在论坛中,突发事件有可能是论坛用户自发讨论的热点话题,更有可能是反复发放的广告主题。本文将基于突发特征的事件发现算法创新性地运用在论坛中,用于发现论坛中的热点话题和广告。实验结果表明,基于用户兴趣模型的用户监测能够有效防范敏感主题的发布,基于事件检测技术的主题监测则可以找出论坛中大部分的广告主题。在两大算法的基础上,我们设计并实现了论坛管理员监测系统,从用户和主题两个角度帮助管理员完成对论坛信息的深入分析。
[Abstract]:With the rapid development of the information age, more and more people like to communicate and share information on the Internet. At the same time, it enriches people's spiritual life. However, because of the freedom, interactivity and openness of the network communication, the network forum has become a high frequency zone of sensitive information and advertising. How to grasp the direction of the public opinion development of the network forum? Mold a clean and active Internet communication platform, become the problem that needs to be solved urgently. The purpose of this paper is to help the forum administrator manage the users and topics in the forum and promote the benign development of the forum. The purpose of user interest model is to discover, extract and express user interest, and show user interest reasonably so that administrator can better understand user's concerns and make corresponding processing. In this paper, a user monitoring method based on user interest model is proposed. By analyzing the user's attention to sensitive topics and user relationship model, the sensitivity of users is defined synthetically, and the potential behavior of users is judged. Emergency refers to the content that is widely spread in a short period of time. In the forum, unexpected events are likely to be a hot topic to be discussed spontaneously by forum users, and more likely to be repeated advertising themes. In this paper, the event discovery algorithm based on burst feature is innovatively applied to the forum to find hot topics and advertisements in the forum. The experimental results show that user monitoring based on user interest model can effectively guard against the release of sensitive topics, while the topic monitoring based on event detection technology can find out most of the advertising topics in the forum. On the basis of the two algorithms, we design and implement the forum administrator monitoring system, which can help the administrator to complete the in-depth analysis of the forum information from the user and subject point of view.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP393.09

【参考文献】