MURRAY, Ky. — Dr. Sean Rife, assistant professor of psychology at Murray State University, worked with colleagues at Verum Analytics, LLC to create a web application that analyzes citations in scientific journals to create numerical indicators representing a paper’s trustworthiness.

Rife, who designed the app’s machine-learning algorithm and wrote the software, noted that scientists use a limited set of terms to refer to other papers in a positive or negative way. For example, the terms “consistent with” and “in line with” are often used to indicate that a publication’s findings have been replicated. On the other hand, terms like “in contrast to” and “unlike” are often used to indicate contrary findings. These commonalities allow a machine-learning algorithm to automatically analyze the text surrounding a scientific citation and classify it as confirming or refuting a scientific claim.

“While there are already a number of ways people measure scientific citations, other metrics simply indicate how many citations an article or author has,” Rife said. “Ours gives some indication of what other researchers are saying about a paper.”

The metric, which Rife and his team call the “R-factor,” is a representation of whether a given scientific paper is regarded favorably or unfavorably by other scientists.

The software is still in its early phases, and Rife indicated that his team is continuing to make it more accurate and user-friendly. They hope to make it a useful tool for researchers, universities and funding agencies. The web app can be accessed at rfactor.verumanalytics.io and was recently featured in Science magazine.

Dr. Sean Rife, assistant professor of psychology at Murray State University, worked with colleagues at Verum Analytics, LLC to create a web application that analyzes citations in scientific journals to create numerical indicators representing a paper’s trustworthiness.
Print Friendly, PDF & Email