【摘要】 随着网络时代的全面到来,如何从大数据的角度去了解和管理网络舆情已经成为一个重要话题。针对这一重要话题,提出一个新的思路:从新闻本身的被关注度和新闻下的评论两方面同时入手,将舆情分析分为两个部分,即"热点事件抓取"和"网民评论情感倾向性分析"。热点事件的抓取通过分布式爬虫实现,网民评论情感分析通过词向量配合训练LSTM神经网络实现。通过这两部分的配合,找出有较高关注度且网民对其有强烈情感的热点事件新闻,从而达到了舆情监控的目的。更多还原

【Abstract】 With the arrival of the Internet era, how to understand and manage network public opinion from the perspective of big data has become an important topic. In response to this important topic, this article opens up a new way of thinking. Starting from the attention of the news itself and the comments under the news, the analysis of the public opinion is divided into two parts, i.e., "hot spot capture" and "emotional polarity analysis of netizens’ comments". The crawling of hot events is realized by distributed crawlers, and the sentiment analysis of netizens’ comments is carried out by using the word vector to train the LSTM neural network. Through the cooperation of these two parts, we can find out the news of hot events with high attention and strong emotional sentiment of netizens which helps to achieve the purpose of public opinion monitoring.更多还原