EXPERTISE AGGREGATION ON BIBLIOGRAPHIC DATA

1
EXPERTISE AGGREGATION ON BIBLIOGRAPHIC DATA

• Presented By
• Brandon Treadway, Delroy Cameron

2
PROJECT PROPOSAL

• Expertise Aggregation
• Bibliographic Data
• Publications
• Publication Impact
• Semantics
• Finding Expertise (Expertise Profiles)?
• Ranking Experts
• Collaboration Networks

3
PROJECT PROPOSAL

• Expertise Aggregation
• Semantic Techniques
• Session Names
• Conference Names
• Paper Abstracts
• Index Terms (Keywords)?
• Problems
• Unreliable DBLP 'ee' links
• Generic Session Names
• Vague/Abstruse Conference Names

4
PROJECT PROPOSAL

• Expertise Aggregation
• Web Crawling
• Collect Paper Abstracts
• PDF Parser
• extract keywords
• abstract, title, index terms

5
IMPLEMENTATION

• Web Crawling
• Yahoo Search API
• Web Search
• Content Analysis
• PDF Parser
• PDFBox
• PDFStripper
• Lucene

6
ARCHITECTURE
Papers-To-Topics
Dataset
PDF Box
PDF Parser
Yahoo Keyword Extractor
Expertise Profiles
Yahoo Web Search API
7
Evaluation

• Subset
• Statistics
• 113 Authors
• 4752 Papers Total
• 3243 Papers Found
• 1509 Not Linked
• 9063 Relationships
• PDF Parser
• 64 Linked by Parser Only (1 )?
• 1445 Not Linked
• 9771 Relationships

8
DEMO

• Web Application
• Keyword Based Search
• Expertise Profiles