https://github.com/greenelab/preprint-similarity-search

A web app that uses machine learning to recommend the most suitable journals based on the text content of your preprint
https://github.com/greenelab/preprint-similarity-search

journals nlp nlp-machine-learning web-app

Last synced: 7 days ago
JSON representation

A web app that uses machine learning to recommend the most suitable journals based on the text content of your preprint

• Host: GitHub
• URL: https://github.com/greenelab/preprint-similarity-search
• Owner: greenelab
• License: other
• Created: 2020-06-16T14:09:22.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2023-05-23T01:19:29.000Z (over 1 year ago)
• Last Synced: 2023-08-21T19:10:49.620Z (about 1 year ago)
• Topics: journals, nlp, nlp-machine-learning, web-app
• Language: Python
• Homepage: https://greenelab.github.io/preprint-similarity-search/
• Size: 43.9 MB
• Stars: 18
• Watchers: 5
• Forks: 4
• Open Issues: 26
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.md

Awesome Lists containing this project

README

        
# Preprint Similarity Search

[⭐ OPEN THE APP](https://greenelab.github.io/preprint-similarity-search/) to start using the tool right away

[📜 READ THE MANUSCRIPT](http://greenelab.github.io/annorxiver_manuscript) for technical details on the machine learning model behind the tool

[🤖 USE THE API](https://api-pss.greenelab.com/doi/10.1101/833400) like `https://api-pss.greenelab.com/doi/YOUR-DOI`

Based on the work and classifiers in the [AnnoRxiver project](http://github.com/greenelab/annorxiver/)

### About

This tool uses a machine learning model trained on 2.3 million [PubMed Central open access documents](https://www.ncbi.nlm.nih.gov/pmc/tools/openftlist/) 

to find similar papers and journals based on the textual content of your [bioRxiv](https://www.biorxiv.org/) or [medRxiv](https://www.medrxiv.org/) preprint.

These results can be used as a starting point when searching for a place to publish your paper.

The tool also provides a "map" of the PubMed Central documents, grouped into bins based on similar textual content, and shows you where your preprint 

falls on the map. Select a square to learn more about the papers in that bin.

The map also incorporates a set of 50 [principal components](https://en.wikipedia.org/wiki/Principal_component_analysis) (PCs) generated from bio/medRxiv.

Each PC represents two high level concepts characterized by keywords of various strengths, illustrated in the word cloud thumbnails above the map.

Select a thumbnail to color the map by that PC.

Deeper orange squares will be papers that correlate more with the orange keywords in the image, and vice versa for blue.