Frida describes her project as the prediction of patient outcomes from Electronic Health Records via machine learning and Natural Language Processing. More specifically, the project aims to develop a decision-support system for detecting early warning signs of cancer among individuals suffering from mental illnesses.

The work will involve developing and applying methods from deep learning, information theory, and NLP to extract and analyse representations of the patient records based on structured clinical data and unstructured text data in the clinical notes.

Frida Hæstrup has a background in Cognitive Science and has been working as a junior developer at Center for Humanities Computing for 3 years prior to starting her PhD, so hers is a familiar face we are very happy to welcome on board once again.