Projects
Cancer, Emotions and LGBTQ+ Cancer Experiences: A Computational Linguistic Study
This research introduces the first annotated corpus of LGBTQ+ forum posts labelled with emotions, used for topic modelling and a hybrid emotion detection method. Addressing gaps in LGBTQ+ cancer care, it explores how forum data can reveal key themes and identify distress signals. Findings highlight eight key topics and a multi-class emotion classifier achieving a 65% micro F1 score, offering insights to improve cancer care for LGBTQ+ patients.
Lived Experience at the Core: A Classification System for Risk-Taking Behaviours in Bipolar
This study explores risk-taking behaviors in individuals with bipolar and their access to support. I conducted semi-structured interviews with 18 individuals with lived experience and 5 healthcare professionals, using content analysis and corpus linguistic methods to develop a classification system of 39 risk-taking behaviours, organised into six domains. Statistical analysis of a Likert-item questionnaire provided further insights. The findings highlight significant gaps in support and the need for better monitoring. The study proposes a standardised classification system to increase awareness around risk-taking behaviours and showcases the utility of computational linguistic methods in health research.
The Hypersexuality in Bipolar - Reddit Corpus (HiB-RC)
This study explores hypersexuality, a common yet under-discussed symptom of bipolar. Using computational linguistic methods, I developed the Hypersexuality in Bipolar Reddit Corpus (HiB-RC) by scraping posts from Redditors who self-reported a bipolar diagnosis. The research uncovers the growing online discussion of hypersexuality, revealing key psychological themes like mood fluctuations, shame, and risky behaviours. I used advanced techniques like corpus-assisted discourse analysis (CADA) and topic modelling to identify 14 thematic concepts, offering critical insights into how hypersexuality is experienced and discussed. The findings highlight the need for better recognition and support in healthcare, urging normalisation of these conversations and training for healthcare professionals to address stigma and improve treatment outcomes.
