Researchers use natural language processing, text mining to enhance PLOS

The Public Library of Science (PLOS) is one of the largest publishers of academic journals and a leading force in promoting open science – free immediate access and unrestricted reuse of scientific and scholarly works of all types.

But with more than 20,000 articles published annually and even more submitted manuscripts, a major operational challenge for PLOS is matching each submitted article with a qualified academic editor. Although PLOS has a network of nearly 6,000 editors to review submitted manuscripts, it can be difficult to assign submissions because editors must be experts on the topic of the article being reviewed.

Capstone posterIn response to these challenges, UVA Data Science Institute Master of Science in Data Science (MSDS) students Jack Prominski and Pragati Shah, led by faculty advisor and DSI professor Rafael Alvarado, elected to work on a capstone project to improve an automated system that matches PLOS article manuscripts with appropriate editors.

“We attempted to solve a key operational challenge and a real pain point in PLOS' review process,” MSDS ‘18 student Jack Prominski said. “We wanted to support an organization on the forefront of the open science movement.”

In 2014, PLOS launched a tool called PLOS Match to aid in connecting manuscripts with appropriate editors. While the system has been effective to an extent, PLOS officials and staff have expressed room for improvement. The team’s research focuses on improving the academic editor acceptance rate, which is defined as the percentage of academic editors who accept the manuscript that PLOS Match has connected them to edit.

They developed a network of more than 300,000 articles that used topic modeling and word embedding features to match editors with relevant work. Most importantly, the team also launched a web application to measure the success of their newly improved program.

The researchers shared their algorithms and results with PLOS, and the publisher will be implementing some of their techniques in day-to-day operations to improve editor matching and acceptance.


J. Prominski, P. Shah, and R. Alvarado. "Editor Matching for Academic Journals Through Rich Semantic Network Development." 2018 Systems and Information Engineering Design Symposium (SIEDS), Charlottesville, VA, 2018.


LEARN ABOUT OTHER CAPSTONE PROJECTS

Completed in:
2018
Project type:
Partner:

Public Library of Science (PLOS)