Aron Henriksson and Maria Skeppstedt presented at the Southern California Workshop on Medical Text Analysis and Visualization during their visit at the Division of Biomedical Informatics, University of California, San Diego. There were many interesting presentations at the workshop, which also led to interesting discussions. The presentation topics most closely related to what is being done in our Clinical Text Mining Group at DSV were “Using Twitter to examine smoking behaviour & tobacco products” by Mike Conway (Division of Biomedical Informatics, UCSD), “Encoding Semantic and Discourse Features for Problem List Generation” by Danielle Mowery (Division of Biomedical Informatics, UCSD) and “Evidence-based Reduction of Perceived and Actual Text Difficulty Using Natural Language Processing” by Gondy Leroy (School of Information Systems & Technology Claremont Graduate University), all of them giving much useful input to our current projects.
Inspiring talks were also given by the other speakers, giving new ideas to text mining and visualisation approaches, e.g. by Andre Skupin (Department of Geography, SDSU) who talked about “Visualizing the topical structure of the medical sciences: A self-organizing
map approach” and Ricky Taira (Department of Radiological Sciences, UCLA) who talked about “Defining and organizing medical NLP tasks”.
Aron talking about automatic synonym extraction using random indexing
Maria talking about features for named entity recognition