Authorship Weightage Algorithm for Academic Publications: A New Calculation and ACES Webserver for Determining...

Area of Science:

• Bibliometrics
• Information Science
• Computational Linguistics

Background:

• Identifying subject matter experts from academic publications is difficult due to jargon and limitations in current search platforms.
• Existing authorship scoring systems often undervalue the contributions of last authors, who are frequently research group leaders.

Purpose of the Study:

• To develop an improved system for identifying and ranking research experts based on academic publications.
• To address the challenges of jargon and inaccurate expertise assessment in publication-based searches.

Main Methods:

• Utilized natural language processing (fastText on Common Crawl) to identify relevant keywords and overcome jargon barriers.
• Implemented a modified authorship positional scoring system that assigns greater weight to the last author.
• Developed a webserver application, APD lab Capability & Expertise Search (ACES), for internal use.

Main Results:

• The ACES system provides a ranked scoring of researchers based on search queries.
• The system effectively incorporates both keyword relevance and a nuanced understanding of authorship positions.
• Successfully addressed limitations in existing expert-finding tools.

Conclusions:

• The ACES system offers a more accurate and effective method for identifying research expertise within an organization.
• This approach enhances the discoverability of researchers by accounting for diverse authorship roles and technical language.