Presented by
Pradyut Kumar Mallick

---

[attachment=11849]
Introduction
Hidden links are ones that real people aren’t supposed to actually notice or click on
Hidden links is a way to guide a search engine to our doorway pages.
New dynamic “hidden link” technique for linking a large highly connected graph in a simple hyperbolic space without cluttering the display.
A cyclic hyperbolic space with hidden links
In a hyperbolic space, the far away nodes/edges (paths) are diminished when the user is not focusing on them.
The user can dynamically warp the display to focus on thousands of different nodes for navigation.
This graph is a non-cyclic hierarchical hyperbolic structure without multiple connected paths
New Technique
The user can easily navigate through all possible paths without tracing many lines and intersections
Robot programs called spiders create search engine databases, computer robot programs that crawl the web seeking search engine content
Pages created as the result of a search are called "dynamically generated" pages .
Definitions
In a directed non-cyclic hierarchical space, there is a primary graph, which links all the nodes in a tree form. These links are primary tree links. The others are non-tree/cross links in a highly connected graph. A node can have one incoming primary page link and many non-tree/cross links.
Definition of Cyclic Hierarchical Space
Primary Path: (tree-link) “AE”
Secondary Path (non-tree/cross link) “AB
Hidden-Link Node
Primary Sub-Space Nodes
Secondary Sub-Space Nodes
Placeholder
Code
The basic page link tag looks something like <a href="hidden.html">click here</a>.
<a href="hidden.html" style="cursor:help">
<a href="hidden.html" style="color:#FF0080">
<a href="hidden.html" style="text-decoration:none">
Algorithm
Build hash table of links in the website.
Partition web log by visitor
For each visitor, partition web log file such that each subsequence terminates in a target page.
For each visitor and target page, find any expected locations for that page:
Hidden Link Applications
CONTENT AND USAGE MINING
CUSTOMER INTERVIEW WEB SERVICE