https://github.com/dimits-ts/disruptive-science-study

Predicting the impact of scientific papers using traditional machine learning models and NLP
https://github.com/dimits-ts/disruptive-science-study

machine natural-language-processing papers prediction-model regression sqlite

Last synced: about 8 hours ago
JSON representation

Predicting the impact of scientific papers using traditional machine learning models and NLP

• Host: GitHub
• URL: https://github.com/dimits-ts/disruptive-science-study
• Owner: dimits-ts
• Created: 2023-04-04T10:16:31.000Z (over 1 year ago)
• Default Branch: master
• Last Pushed: 2023-12-11T19:56:22.000Z (11 months ago)
• Last Synced: 2024-04-22T02:45:09.617Z (7 months ago)
• Topics: machine, natural-language-processing, papers, prediction-model, regression, sqlite
• Language: Jupyter Notebook
• Homepage:
• Size: 179 KB
• Stars: 1
• Watchers: 1
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# disruptive-science-study

A [recent paper in Nature](https://www.nature.com/articles/s41586-022-05543-x) created a stir in the scientific community, arguing that science is becoming less disruptive over time. According to the study, there are fewer groundbreaking papers in recent years. It appears that trailblazers are rare and that most research tends to build and expand existing research rather than opening up new paths of inquiry.

While there already is a method with which we can judge whether a paper was impactful, there hasn't so far been an attempt to *predict* its impact. This project aims at achieving that using publically available data, traditional machine learning models and modern NLP methods.

Credit to professor Panagiotis Louridas for originally assigning the project, as well as migrating the data to an easy-to-use SQLite DB.