Intranet Page Ranking via Workstation Content Mining
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing page-ranking systems for intranets are ineffective due to low density of hyperlink references and lack of economic incentives for cross-referencing valuable pages, unlike the Internet, where PageRank algorithms rely heavily on external hyperlink counts.
Innovation Solution
A page-ranking method and system that mines user workstation content, such as browser bookmarks, e-mail, and text files, to detect and rank page references, providing an alternative voting-based system that considers content relevance and expertise within the network environment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If hyperlink reference density is used for page ranking on intranets, then the ranking system can be implemented with existing PageRank algorithms, but the ranking quality deteriorates due to low reference density and lack of economic incentives for cross-referencing
Solution Approach 1:
The patent introduces an intermediary component (the mining module) that extracts reference information from user workstations. This intermediary bridges the gap between the lack of natural hyperlinks on intranets and the need for reference data, collecting references from bookmarks, emails, and other user-generated content to enable effective page ranking without relying on dense hyperlink structures
Solution Approach 2:
The patent copies the effective mechanism from Internet search engines (hyperlink-based PageRank) and adapts it to intranet environments by mining user workstation content for reference information. Instead of relying on actual hyperlinks between intranet pages, the system copies the ranking methodology and applies it to extracted reference data from user devices, effectively transferring the successful Internet approach to the intranet context
2Measurement precision
If user workstation content is mined to detect page references, then search relevance improves, but system complexity increases due to additional mining and processing modules
Solution Approach 1:
The mining module is designed with multi-functionality, serving multiple purposes: it extracts page references from various sources (bookmarks, emails, documents), collects user expertise information, and provides data for both page ranking and user profiling. This universal component reduces overall system complexity by consolidating multiple functions into a single module rather than requiring separate systems for each function
Solution Approach 2:
The system leverages existing user-generated content on workstations (bookmarks, emails, documents) that users have already created and organized. By mining this pre-existing self-service data, the system avoids the complexity of creating new reference mechanisms or manual annotation processes, instead utilizing content that users have already produced for their own purposes
Data Source
AI summary
A page-ranking method includes mining a portion of content of a user workstation which is connectable to a network to detect references to pages of the network. The pages may be ranked based on the detected references.