面向复杂决策场景的认知图谱构建与分析

Construction and Analysis of Cognitive Graphs for Complex Decision-making Scenarios

在复杂网络场景中有效地甄别不同类别事件的性质和影响、辅助个体和组织决策、避免决策偏误,是复杂社会系统管理的重要研究问题。本文基于海量网络用户的交互内容数据,对网络数据中隐含的个体认知信息进行深度分析和系统化建模,提出了面向复杂决策场景的认知图谱构建与分析方法,从而对公众情绪的演变和群体性事件的走向进行可靠的预测。在中国股市中的高估信息、低估信息和披露信息3种不同类型事件中进行实验验证,结果表明:本文提出的认知图谱构建与分析方法能够解析不同类型复杂场景中个体认知要素变化与事件时空演化态势的关联关系,并能够有效支撑公众情绪的演化和群体性事件的预测与分析。

How to effectively identify the nature and influence of different types of events in complex network scenarios,assist individuals and organizations in decision-making,and avoid decision-making biases are important research issues for the management of complex social systems.Based on the interactive content data of massive online users,this paper conducts in-depth analysis and systematic modeling of the hidden individual cognitive information in cyberspace,and proposes the cognitive graph construction and analysis method for complex decision-making scenarios.It can reliably predict the evolution of public sentiment and the trend of group events.The experimental verification results in three different types of typical cases,including overvalued information,undervalued information and disclosed information in the Chinese stock market,show that the cognitive graph construction and analysis method can effectively analyze the underlying relations between the individual cognitive elements and the spatial-temporal evolution of events in different types of complex scenarios,which can provide significant insights for us into predicting the dynamic changes of the public sentiment and the mass events.