中文摘要：

网络已成为当今世界重要的信息载体，但是网络信息良莠不齐，对人们的生活造成了很多负面影响，因此，如何正确识别网络中的敏感话题，是当前网络舆情分析与监管的重要任务之一。本文以识别网络论坛中的敏感话题为目标，基于网络论坛文本在结构和表达上表现出的篇幅短、结构不完整、文字口语化等特性，将该类文本表示成基于向量空间模型的文本矩阵，并根据网络敏感话题具有先验知识和态度倾向性等特点，提出了基于倾向性词典的特征提取方法，可有效提高网络敏感话题识别的正确率，最后通过实验验证了这一改进的有效性，证实了本文的研究价值。

英文摘要：

Internet has become an important platform for information spreading, but not all of the information on the Internet is useful to people＇s daily life. Therefore, how to correctly detect sensitive topics on the Internet is one of the most important tasks for network public opinion supervision. This paper aims to detect sensitive topics in online communities on the basis of the features of Internet texts, such as short length, incomplete structure and colloquialism, using the vector space model to represent the texts. We propose a new feature extraction method based on the sentiment lexicon to improve Internet sensitive topic detection accuracy. Finally, the effectiveness of our method is discussed based on data from experiments.