Skip to content

cjbayesian/ml4h_paper_2019

Repository files navigation

ML4H 2018 NeurIPS paper analysis

Pipeline:

Using anaconda:

conda env create -f environment.yml
source activate ml4h_papers
jupyter notebook
  1. Get the metadata for all papers NeurIPS_Proceedings_Metadata.ipynb
  2. Download the pdfs NeurIPS_Proceedings_dl_pdfs.ipynb
  3. Extract raw text from pdfs:
python pdf2text.py ./data/ML4H2018/pdf/ ./data/ML4H2018/txt/
python pdf2text.py ./data/NeurIPS2018/pdf/ ./data/NeurIPS2018/txt/
  1. Pre-process raw text
    • Run text pre-processing NeurIPS_Proceedings_preprocess.ipynb OR
    • Run the create_r_dataset.R script to generate .rdata files
  2. Optional: run the example topic model and visualization script ldavis.R.

Requires

Python

  • Python >= 3.7
  • requests
  • beautifulsoup4
  • pandas
  • json

R

  • tm
  • LDAvis
  • lda

Other

  • pdftotext version 0.64.0

About

Data Pipeline for analysis of ML4H 2018 workshop at NeurIPS.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published