Automating HEAL Grant Data Characteristics using NLP/ML

• HEAL seeks to improve both pain management and prevention tactics for opioid use disorder.
• We use supervised machine learning and natural language processing methods to help automate HEAL grant categorization.
• Automating classification of HEAL awards for portfolio analysis will:
  • Significantly reduce the time burden of portfolio analysts within HEAL.
  • Highlight research themes, connect investigators studying aligned targets and interventions and determine promising areas for allocating research support.

Set up and activate the conda environment by running the following lines:

conda env create -f new_environment_1.yml
conda activate new_environment_1

Structure of Repo:

• RCDC - includes all RCDC files of term sets we used from them.
• project_info - includes slide deck of the project.
• results/ - all results and figures are saved here. Information about these folders are contained in their readmes.
• src/ - all coding files are saved here. Instructions to run the files to replicate results are contained in code commentation. Methods included rule-based Natural Language Processing approaches, as well as supervised Machine Learning text classification models ex. Random Forest, K-Nearest Neighbors, Support Vector Machine, Logistic Regression. The README.md file inside this folder contains detailed description of the work done and the limitations thus far.
• term_sets/ - includes all the term sets we are currently leveraging and words we discovered using unsupervised learning methods.
• I couldn't push the data to the repo due to security reasons, but a lot of my argparse functions refer to the file paths I stored it in on my local drive. I had an original_data folder, as well as a cleaned_data folder for my data.

Learn more about HEAL's mission at their website: