Laying the Explorative Groundwork for Document Quality Assessment and Perspective Detection

In this abstract, we address the problem of information overload that web users face when researching controversial topics on the web. The ultimate goal of our project is to design a tool to aid users in reviewing and learning about the quality and content of large document collections online.

We highlight the challenges as well as the previous and future steps taken to develop a document information quality visualization tool. This lays the explorative groundwork for developing a web browser tool which efciently informs users about the quality of web documents according to 8 quality dimensions. We also address the challenges of identifying author and source perspectives in documents and collections of controversial debate related text. We use Natural Language Processing (NLP) methods to explore these linguistic features and to extract important textual content with which to contextualize the document information quality. Finally, we motivate our planned crowd-sourced study which is designed to explore the feasibility of automatically evaluating contradictory perspectives across controversial documents. We take the highly publicized vaccination debate as a focal point with which to focus development