移动社交信息智能态势感知分析

Analysis of intelligent situation awareness of SNS

随着微博、社交网站、移动互联网的快速兴起,网民通过信息网络获取信息、表达诉求和公共参与日益增多,舆情监测和引导面临的挑战日益严峻.针对移动社交网络环境下舆情态势感知存在的问题,本文提出了智能态势感知方法.智能态势感知结合了层级时序记忆,提取的信息流特征更适合于不确定性高的无标度信息动态环境,能够有效降低历史虚警数据的干扰;基于自学习的鲁棒特征筛选方法,实现了对提取节点数据特征的自动筛选,并在部署移动社交网络服务器应用的环境中以及无人值守的业务情景中实现了态势感知数据的自动汇聚;构建了态势知识库,使态势特征更加鲁棒,不因小世界网络拓扑的动态改变而丢失特征或造成虚警,使态势感知更适用于小世界网络环境;提出了空检验矩阵,对态势可视化结果进行精练并剔除错误匹配点,具有加速态势可视化的特点.随着移动社交网络环境的发展成熟和网络用户的成倍增长,智能态势感知为进一步提升应急信息平台的智能化程度、实现有效的舆情监测提供保障.

With the rapid rise of microblogging, social networking sites, and mobile Internet, users＇ access to information, public participation, and expression of their demands is increasing significantly. Thus, the challenges faced by public opinion monitoring and guidance are becoming increasingly serious. To address the problems of mobile social network public opinion monitoring, this paper presents an intelligent situation awareness analysis. Intelligent situation awareness combines with the timing of memory hierarchy, extracting information flow char- acteristics better suited to a highly uncertain, scale-free dynamic information environment, effectively reducing the interference of false alarm history data; is based on robust self-learning screening methods to achieve au- tomatic screening of the node data extraction features and to facilitate the automatic aggregation of situation awareness data in mobile social networking application server environments and unattended scenarios; builds situation knowledge to improve robustness and not lose the characteristics or cause false alarms because of dy- namic changes in the network topology of small-world, making situation awareness more suitable for small-world network environments; makes empty inspection matrix scouring the results and eliminating false match points to accelerate visualization. With the maturing of the mobile social networking environment and doubling of the number of users, intelligent situation awareness provides protection to further enhance the level of intelligence of the emergency information platform and facilitate effective monitoring public opinion.