Research On Event Visual Analytics Based On Knowledge Graph

News event on the Web has become one of the main research areas in artificial intelligence.Event contains rich time and spatial information and connects to each other.force-directed layout based graph visu-alization technique hinders the cognition process as positions of the nodes are randomly generated.Morever,data overload problem caused by visualizing the whole data in node-link style graph layout block relation analysis.Therefore,how to design an effective method to analysis event is a very meaningful study topic.To solve above problems,this thesis focuses on applying data visualization techniques based on different type of data to event visual analysis.Specifically,three main contributions of the thesis are as follows:1.A method to distinct the importance of the entities in event was proposed,data overload problem could be overcomed by exploring interestness data layer-by-layer.2.A hierarchical graph algorithm in which nodes are ordered by date was proposed.The algorithm we proposed firstly ordered entities involved in event by its date of occurrence,and then applied the parent-centered hierarchy graph layout algorithm to visualize the data.3.The proposed system was designed and implemented,and the effectiveness and usefulness of the system was verified with case study named "ofo shared bike" and user test separately.