Massive Questionnaire Analysis and Sleep Apnea
Adam Paul Wisniewski · TSpace (University of Toronto) · 2015
Massive data analysis is a new and growing field in applied mathematics. Questionnaire datasets are ubiquitous in many different fields and they usually contain massive structures. This study surveys prominent supervised and unsupervised massive data analysis techniques for classification and clustering, including k-means clustering, spectral clustering, hierarchical clustering, support vector machine, and random forest. It also explores a novel new approach to questionnaire data analysis developed by Professor Ronald Coifman that incorporates the dual geometry metric. All of these techniques are then applied to an easily obtainable and noninvasive questionnaire dataset in order to predict the severity of a patient's obstructive sleep apnea. The resulting analysis is found to have significant predicative power for female patients.