Information De-identification from Free-text Questionnaire

Title: A tool for the removal of identifying information from free-text fields in questionnaire data

The project addresses the challenge of identifying and obfuscating personal health information (PHI) in open-ended survey responses. It proposes a solution using Named Entity Recognition (NER) methods from Natural Language Processing (NLP). Challenges include the lack of annotated data and context in survey responses. The proposed model demonstrates improved performance compared to the baseline method. Future directions involve obtaining more labelled data, incorporating survey questions for context, and utilizing advanced NLP models. The document emphasizes the ongoing academic inquiry into obfuscating PHI in survey data.