/Biorec

Primary LanguagePython

Biorec: Conference and Journal Recommendation App

Open in Streamlit Heroku

Table of Contents

  1. Background
  2. Layout
    1. Home Page
    2. Side Bar
  3. Usage
    1. Text Guide
    2. GIF Guide
  4. Notes
  5. Acknowledgements

Background

Biorec was developed to simplify the publishing lives of academics.

In the present day there is an astronomical amount of conferences to submit to and no user-friendly way of parsing which conferences would be most relevant for any given publication/poster/abstract.

The goal of this conference recommender is to accept abstracts or resumes (eventually) as input and to recommend relevant conference series that would be good avenues for publications on topics similar to the input.

Layout

Here's what the home page of the app looks like:

Home Page

Home Page

You have two optional expanders:

Raw Corpus Preview: Shows the raw corpus preview shows the raw, unedited data that the recommender uses for recommendations; is human readable.



Raw Corpus Preview

Tokenized Corpus Preview: Shows the tokenized "soup" of data that the recommendation algorithms utilize; the soup is not human readable.



Raw Corpus Preview

Side Bar

The side bar contains input and output options for Biorec.

Biorec can provide recommendation using 4 different algorithms:

The user chooses the format in which the query is entered in to the recommender. Currently available choices are as a text file (.txt extension only) or as a plaintext query into a text box.

The user also selects the number of recommendations to recieve, up to a maximum of 50 recommendations. The user can type out the exact number of recommendations required, or use the plus/minus buttons to increment or decrement the counter by steps of 5.

Usage

Text Guide

Use the recommender in 8 easy steps:

  1. Select the recommendation algorithm → The default recommendation algorithm is BM25

  2. Select the input method → The default input format is text file

  3. Enter the number of desired recommendations → The default number of recommendations is 10

  4. Upon the completion of step 3 press the "Submit" button located below the number of recommendations

  5. Enter the query in the query input option available → The default input option is file

    If you selected file you should see the following file picker

    file picker

    If you selected text box you should see the following text box (press Cntrl + Enter to submit text box query)

    file picker

  6. View your recommendations

  7. Choose if you want to download the recommendations (Download only available in the form of a csv file) → The default is Yes

  8. Repeat steps 1-7 for as many permutations as desired

GIF Guide

Here I demonstrate how to use the Conference Recommendation App with the Doc2Vec algorithm and save the resulting recommendations.

gif demo

Notes

To reset the app press: Ctrl+Shift+R

For reviewers:

  1. The downloaded csv file will contain a column with the title "Rating"
    1. Please use that column to provide ratings for the recommendations proivded on a scale from 1 to 5
      1. Where 1 indicates that the recommendation was not pertinent to your query and 5 indicates that the recommendation was completely pertinent to your query
  2. Please give ratings for each recommendation algorithm

Acknowledgements