Jointly Event Extraction And Visualization On Twitter

As a representative media of the new era,Micro-blog has recently become a popular platform for expressing ideas,sharing information and exchanging views on the network,which has a great impact on society.Event extraction,one of the core problems in the field of information extraction,aims to extract information that people are interested in from unstructured text containing events and express them with structured form instead of natural language,such as what happened,to whom,where,and when.It has dramatically practical significance and great applicant value to study the technology of event extraction on Micro blog.Text visualization is one of the most important tasks in data visualization.Its main purpose is to use rich graph and image to reveal the hidden information from text data.The technology of text visualization can present a high level overview of the core context of text information,as well as display them intuitively.Therefore,jointly event extraction and visualization on micro blog text has great value applicant to revel latent event information and relationship between them.So here we focus on jointly event extraction and visualization on Micro-blog.Our main contributions are summarized as follows.1.We make a research on the technology of event extraction and visualization and propose an event extraction and visualization approach based on Latent Event Extraction&Visualization(LEEV)Model.LEEV is an unsupervised Bayesian latent variable model.It is the extension of the Latent Event Model(LEM)by incorporating the coordinate information.This paper describes the system framework,the process of the extraction and visualization,the description of LEEV model and the parameter estimation method in detail.Two datasets containing 2453 and 1000 tweets are used to evaluate the effectiveness of our approach.We compare the extraction and visualization performance of our model with the state-of-the-art approach.The results of our method on two datasets both outperform the state-of-the-art approach.2.Noticing the fact that the intrinsic geometry of textual data is a low-rank,non-liner manifold lying in the high dimensional space,we improve the original model and propose an event extraction approach based on LEEV+R.We modify the LEEV model by putting an additional manifold regulation to it.Same datasets as LEEV model are used to evaluate the effectiveness of the LEEV+R model.The experimental results show that LEEV+R model outperform LEEV model on both of the two datasets.This paper consists of five chapters.The first chapter introduces the research background and significance,the motivation and the main research content.The second chapter describes the related theories and existing technologies of event extraction in Twitter and text visualization.The third chapter introduces the proposed approach based on LEEV and related experiment.The fourth chapter introduces the proposed approach based on LEEV+R and related experiment.The fifth chapter is the summary and future outline of this work.