In this appendix we give detailed examples of creating graphs from web documents. We show several web documents, including their HTML source and view when rendered in a web browser. We also show the resulting graphs that can be created from the web page content. The documents, which are taken from the University of South Florida web site, are:

1. Document 1: the section of the undergraduate catalog pertaining to the College of Engineering (http://www.ugs.usf.edu/catalogs/0304/enggen.htm)

2. Document 2: a page discussing dining and meal plans at the university (http://www.usfdiningservices.com/mealplans.php)

3. Document 3: a page describing safety issues for students studying abroad (http://web.usf.edu/iac/studyabroad/safety.htm)

Document 1, as it appears when rendered in a web browser, is given in Figs. A.1 A.8. The original HTML source code is presented in Figs. A.9 A.15. We have created a graph from this document, using the standard representation (Sec. 3.2). We did not do any dimensionality reduction by removing nodes. This graph was then exported to DOT format and rendered using GraphViz. ^[1] The resulting graph is too large and complex to display easily, but we show the largest connected component in Fig. A.16. Nodes with large degrees include: ENGINEER, COLLEGE, STUDENT, PROGRAM, SYSTEM, DESIGN, and COMPUTER. Because the individual nodes can not be seen clearly, we have provided close-ups of sections of Fig. A.16 in Figs. A.17 A.32. To accomplish this, we have divided Fig. A.16 into sixteen equal sections using a grid. We refer to each section as ( r, c), where r gives the row...