Research And Application Of Visualization Guided Document Reading

In recent years,due to the growing number of publications,researchers have to read much more literatures.Therefore,how to read a scientific article in an efficient way becomes an importance issue.However,a scientific article is usually the essence of an academic study,which covers many of the arguments and findings.This makes it very difficult for readers to get the core argument of an article in a short time.Also,when reading an article,it's necessary to read its references in order to get a better understanding.However,the citation network of an article is complex structured.Seeking for a relevant reference in a citation network is easy to get lost in this complex network of literatures.So how to differentiate between the relevant and non-relevant references,and how to avoid being lost in a large document collection are also noteworthy issues.In this paper,we use visual text analysis techniques to study the scientific literature reading.In order to solve these problems,this paper proposes a new method based on the reading objective-based text summarization technology to extract the key sentences and LDA(Dirichlet Allocation Latent)is used to analyze the content of scientific literature which are based on the research on the relationship and characteristics of the scientific literature in the citation network and document analysis.In addition,we present GUDOR(GUided DOcument Reader),a visualization guided reader based on citation and summarization.It(1)extracts the important sentences from a scientific article with the objective-based summarization technique,and visualizes the extraction results by a multi-resolution method,in order to facilitate readers to locate the core content of the article(2)identifies the main topics of the references with the topic model,and designs a variety of visualization techniques,including word cloud,tree map,radial diagram,etc.,to show the core content of the reference and the relationship with the article;(3)tracks user's behavior in the whole reading process to keep him or her focusing on the reading objective and prevent be lost.In addition,we describe the functions and interactions of our system in a usage scenario and validate its utility by a user study which turns out that the proposed system has good scalability and good user experience.Finally,in the case study,we analyze the reading behavior of different users and obtain many different reading patterns.In the future,we will model and analyze the user reading of literatures in order to facilitate understanding of literatures as the recommendation.