Undergraduate Projects

This project is a currently in-progress development for a Natural Language Processing tool, designed to extract meaning from (migraine) clinical patient records.The tool is being developed in Python through the combination and comparison of existing packages, toolkits and API’s to extract meaningful patient information, and reorganise it into a more succinct format. The primary goal is to be able to present patient data accurately and easily for clinical practice purposes, with the additional goal of allowing further exploration of patient data to better understand illness progression and treatment efficacy in relation to patient characteristics. Current progress has consisted of a literature review, selection of key NLP tasks and metrics to score performance of these tasks, and exploration of a number of tools to determine current functionality for clinical Natural Language Processing. At present, medical entity and relational information can be extracted from raw text with high accuracy, and a number of options have been set forth to improve benchmarking, and allow the tool to extract temporal information which is integral to text meaning.