Research On Event Analysis And Visualization In Email

Filtering and analyzing Email texts can effectively assist people in understanding important event information and human interaction behaviors that occurred in the past.The rich text,images and attachment information in the Email datasets also has great potential for mining.Most of the existing Email visualization researches mainly conduct related analysis by showing the relationship between Email subject and time,Email visualization element design,or visual Email social network.Among them,the subject filtering and clustering methods for Email datasets are relatively simple,and the effect of extracting Email event sequence information is not significant.A large number of Email with low analytical value will also bring visual interference to people when reviewing history,which will consume people more energy.The improvement of the above problems is of great significance for reviewing important past events,improving team collaboration relationships,and improving collaborative work efficiency.Therefore,this thesis focuses on Email datasets cleaning and mining methods and event sequence analysis and visualization methods.The main research work of this thesis is summarized as follows:(1)Summarize and analyze Email social networks and event sequence visualization methods: First,the visual element design,metaphor patterns and related analysis processes of Email social networks are introduced,and then the common algorithms and features of text mining are summarized.Finally,the mining,analysis steps and visualization methods of event sequence are summarized.(2)Propose a filtering analysis model for Email datasets and use visualization to verify it: First,establishes a filtering model for Email datasets,and designs corresponding visualization tasks to clean and analyze the text of Email datasets from four dimensions,namely Spam filtering,Email subject keyword extraction,Email flow characteristic analysis,and business field change analysis.Then design a combination of multiple visual views to help people understand the content of the Email datasets more clearly.(3)The event sequence mining and analysis model in Email is proposed and realized by visualization: First,the concept of Email thread and the basic visual element metaphors of event sequence are defined,and then a combination of nodes and lines is used to visualize Email.The event sequence information of social networks explores the characteristics of the event sequence between different months,and analyzes the important events,redundant processes,abnormal problems,project collaboration,etc.,and proposes improvement suggestions.The user study results further prove the practicality and effectiveness of the visualization method in this thesis.