中文摘要：

互联网时代，网络舆情的庞大数据规模和舆情分析的计算复杂性，使对网络舆情的分析和实时掌控变得愈发困难。面向快速、不断产生的网络舆情采用流式计算进行实时处理的分析模型，在时效性、突发性和无限性三个方面都更加符合网络舆情的自身特性。基于流式计算的网络舆情分析模型分为数据收集、舆情分析和舆情治理三个部分，通过对语义保障和负载控制等关键技术的把控，可以实现个案把握向整体掌控、被动响应向主动分析的转变。基于流式计算的网络舆情分析模型具有可扩展性，能够联合众多服务器及资源，具有平台优势，能够解决地方舆情分析中面临的技术门槛，保障网络舆情分析的准确性与及时性。

英文摘要：

During the Internet age, the network public opinion analysis and real-time control are becoming more difficult since the large data scale of network public opinion and the computational complexity of public opinion analysis. For rapid, stream data continuously generated in real-time processing, the analysis model has three aspects of advantages including timeliness, sudden and unlimited which is more in line with its own characteristics. The model of network public opinion analysis based on stream computing can be divided into three parts ： data collection, analysis of public opinion and public opinion management, through the key technologies such as semantic security and load control, this model has also realized transformation from the case to overall control and the passive response to proactive analysis. The model of network public opinion analysis based on stream computing is scalable, can be combined with many servers and resources, with the advantages of the platform, it is possible to solve the technical barriers faced by local public opinion analysis to ensure network public opinion analysis accurate and timeliness.