Nov 16, 2017
Jolley Hall, Room 309
"Instructive Learning: A Framework for Learning from Clinical Big Data"
Adviser: Yixin Chen
The amount of data being digitally collected and is vast and expanding rapidly. As a result, the data science is also advancing to enable organizations to convert this vast resource into information and knowledge that helps them achieve their objectives. Big Data is invented to describe this evolving technology. By far, big data has been successfully used in astronomy, retail sales, search engines and politics. Big data analytics frequently exploits analytic methods developed in data mining, including classification, clustering, and regression. In this work, we focus on clinical big data applications. Big data technology has many applications in clinical area, such as predictive modeling and clinical decision support, disease or safety surveillance and research. However, clinical big data faces challenges such as hard to access, limited labeled data, mixed categorical and numerical features, mixed static and time series data. In this work, we propose to develop an instructive learning framework to effectively learn from clinical data while overcoming these challenges. Instructive learning consists of four key components: a preprocessing module, an unsupervised learning module, an experts analytic module, and a feedback module. I propose to develop the instructive learning framework and evaluate its performance in real clinical applications.