Vendredi 22 novembre 2024
Natural Language Processing Methods for Population Health Research
Dan Lizotte
Professor, Western University
Heure: 13h30
Local: PLT-2744
Résumé: Natural Language Processing methods offer a way of analyzing text data that traditionally would be used for qualitative research, for example using ethnographic or grounded theory methods. They offer opportunities both to scale up analyses to larger datasets than are feasible for humans to analyze and to scale out analyses to learn about and support populations that may otherwise be inaccessible because of stigma. In both cases, natural language methods offer new possibilities for better understanding population health. In this presentation, I will describe projects spanning biostatistics, computer science, and health research that apply natural language methods to population health questions, explain the technical and conceptual challenges we needed to overcome, and relay what we learned. I will end by proposing some future directions at the intersection of these disciplines.
Content advisory: Topics of mental health, self-harm, substance use, and homelessness will be discussed.
Biographie: Dr. Dan Lizotte is Associate Professor in the Department of Computer Science and the Department of Epidemiology and Biostatistics at Western. His research aims to support health decision-making by developing and applying machine learning and statistical tools to new sources of data including electronic health records and social media to better support patients and health professionals, particularly in public health and primary health care. His methodological research combines machine learning, optimal sequential decision-making, and multiple objective optimization. He is also the director of the Rotman Institute of Philosophy at Western and he is currently working with the Alliance for Healthier Communities to develop tools for research and decision support.
Note: La présentation sera donnée en anglais.
http://www2.ift.ulaval.ca/~quimper/Seminaires/
