CLIS Home > Research & Projects > Seminar Series

CLIS Research Seminar Series

February 08, 2008 @ 10-11 Am
"Explaining and Predicting Web Page Status"

Padmini Srinivasan, Ph.D.
The University of Iowa

Friday, February 8, 2008, 10-11 am
Room 2119 Hornbake Building, South Wing

Abstract

In-links to a web page determine its visibility, accessibility through surfing, accessibility by search engine crawlers, position in search engine rankings and more generally its popularity or status.  By linking to a web page an author acknowledges that the target page has had some influence on the intent of the page being built.  As status typically requires time to develop, new pages face a significant hurdle particularly since the web reveals a “success-begets-success” nature.  A challenge that is both of inherent interest and of much practical value is to be able to predict the future status of a new web page.  

In this research we present a systematic study of a variety of features in terms of the capacity to explain and predict page status. It may be noted that although the web is replete with experience-based guidelines on how to gain visibility for pages, few systematic studies explore the key underlying page properties.  We are particularly interested in features that are under the control of the page designer although we also control for other features in our study.  Examples of features we study include information volume, specificity and location.  Still others such as information brokerage are motivated by research in sociology relating to how humans garner status in societies.  In summary, our results indicate that the features we study can be used to explain and predict status, especially when we take a topical perspective on the problem. The talk will motivate this research, present our experiments and offer a discussion of key findings.

This research has been conducted in collaboration with Gautam Pant, Assistant Professor, School of Accounting and Information Systems, David Eccles School of Business, The University of Utah.

About the speaker

Padmini Srinivasan received her Ph.D. in Information Science from Syracuse University. She has been a professor at the University of Iowa since 1989. She holds a joint appointment in the School of Library & Information Science and the Department of Management Sciences with invited appointments in Computer Science and the College of Nursing. Dr. Srinivasan’s research interests are in core information retrieval problems such as query expansion, retrieval and ranking algorithms. She is also interested in studying web-based phenomenon such as the design of web crawlers; in designing algorithms for mining hypotheses from text; and more generally in building text-based applications in domains such as biomedicine.