Graphical Analysis of Electronic Medical Records

Aryya Gangopadhyay, Department of Information Systems, UMBC
Rose Yesha, Department of Information Systems, UMBC

The goal of this project is to analyze clinical charts that are part of electronic medical records (EMRs). The proposed method will automatically discover latent thematic patterns from the clinical charts. The themes discovered will be presented to end users using visualizations that can subsequently be linked back to the relevant text segments of the corpus. A search function will be provided to look for information about patterns specified by the users. Our algorithms will be compared with other methods such as topic models and validated with ground truth information. The scalability of the methods will be tested by measuring execution time against data with varying sizes.