B1325
Title: Semi-supervised learning to predict adherence to psychotherapy with mHealth data
Authors: Samprit Banerjee - Cornell University (United States) [presenting]
Abstract: Smartphones provide an interactive interface that can passively measure various aspects of the users' behaviour from device sensors and actively collect self-ratings (e.g., mood, stress, etc.) obtained via daily ecological momentary assessment. Taken together with traditional clinical assessments, these measures have the potential to provide unique insight into the treatment trajectories of patients with major depressive disorder undergoing psychotherapeutic treatment. Specifically, patient adherence to psychotherapy sessions is a necessary first step to assess barriers to adherence and personalize future sessions in order to improve adherence and, therefore, efficacy. Such predictions have unique challenges due to the noisy nature (missing or under-reporting) of passive and active mHealth data. The nature of missing passive data is unique in the sense that the missed labels are not observed. These and other challenges of mHealth data analysis are introduced, and semi-supervised machine learning algorithms are proposed to address these challenges.
