基于大数据和人工智能进行网络舆情分析的研究

Research on Network Public Opinion Analysis Based on Big Data and Artificial Intelligence

随着互联网、移动互联网、物联网、社交网络等技术和应用的兴起,媒体技术的革命正在造就一个全新的舆论环境,网上言论已达到前所未有的活跃程度,互联网日益成为社会各阶层利益表达、情感宣泄和思想碰撞的平台,进而产生巨大的舆论信息。面对网络上产生的海量信息数据,快速筛选出有用的网络舆情信息,通过网络舆情分析、监控民情意见、情感倾向,为相关部门提供及时的协助决策和分析结果,快速形成处理网络上突发性群体事件的可行性方案,是保障大数据舆论监督有效性的关键。文章提出了一种基于大数据云计算、信息预处理优化聚类算法及中文NLP(自然语言处理)情感倾向分析算法的人工智能网络舆情分析平台。加快有效信息的筛选速度及民情导向的分析速度,保证在海量网络数据的环境下,舆论监控工作的及时性和有效性。最后通过实验,与传统的统计式大数据信息分析系统进行比较,该方法具有信息收敛速度快、信息分析高效,可靠性高,特别是在做好重点关注领域的分类训练后,随着采集数据量的增长,对舆情导向分析结果也更准确。

With the rise of internet the Internet, mobile internet, internet of things, social network and other technologies and applications, the revolution of media technology is creating a brand-new public opinion environment. 3 Online speech has reached an unprecedented level of activity, and the internet has increasingly become a platform for all social strata to express their interests, vent their emotions and collide their thoughts, thus generating huge public opinion information. Facing the massive information data generated on the network, it is the key to ensure the effectiveness of public opinion supervision of big data by quickly screening out useful network public opinion information, providing timely decision-making assistance and analysis results for relevant departments through network public opinion analysis, monitoring public opinion opinions and emotional tendencies,and quickly forming a feasible scheme to deal with sudden group events on the network. This paper proposes an artificial intelligence network public opinion analysis platform based on big data cloud computing, information preprocessing optimization clustering algorithm and Chinese NLP(Natural Language Processing) emotion tendency analysis algorithm. 6 Accelerate the screening speed of effective information and the analysis speed of public sentiment, and ensure the timeliness and effectiveness of public opinion monitoring under the environment of massive network data. 7 Finally, through experiments, compared with the traditional statistical big data information analysis system, this method has the advantages of fast information convergence, high efficiency of information analysis and high reliability. 8 Especially after the classification training of key areas of concern, with the increase of collected data, the results of public opinion-oriented analysis are more accurate.