Faculty, staff, fellows and students are welcome to hear Dr. Varun Chandola, Assistant Professor in Computer Science and Engineering and the Center for Computational and Data-Enabled Science and Engineering, discuss several machine learning based approaches for patient sub-typing from temporal clinical records. Many of these methods fall under the task of data clustering. However, owing to the complexity of the data and the inherent temporal dynamics, standard clustering methods are not directly applicable. He will discuss an extension of the widely used K-means clustering algorithm that we have developed to handle such data. Part of this work was completed in collaboration with Dr. Chet Fox from the Department of Medicine. Dr. Chandola will also discuss other probabilistic models that have been recently proposed in the same context. He will discuss the strengths and weaknesses of these methods in understanding Chronic Kidney Disease (CKD) using clinical data available from the DARTNet Institute. He will also briefly talk about new results on another freely accessible critical care data set, the MIMIC-III database.
Cheryl Kennedy
Email: clk2@buffalo.edu
Phone: 7168884847