MedVector

Screenshot of the MedVector program

Inspiration

We wanted to help clean up the inefficiencies in hospitals, which potentially impact people's lives. We wanted to help make sure people got the best care possible at hospitals by helping mitigate delays.

What it does

This an Eclipse-based program that looks for and matches similar medical records using a weighted term vector algorithm, groups them into blocks, and lets the user browse the data according to how closely data pairs match.

How we built it

This program was built in Eclipse using Java and Apache libraries. The GUI was built using Processing libraries. We used Febrl, available at link to randomly generate sample medical records.

Challenges we ran into

We quickly realized that the algorithm had a very slow runtime, and we helped improve this by implementing phonetic algorithms and by taking advantage of blocking.

Accomplishments that we're proud of

We cut the problem down into steps and ended up creating a nice UI for the project.

What we learned

We learned more about how to handle large amounts of data and how to split up tasks to multiple people.

What's next for MedVector

This algorithm gets improved for more efficiency and gets used for even larger amounts of data. In terms of this submission, the focus was mostly placed on the implementation of the algorithm. Future distributions may have more work on the usability, flexibility, number of features, and user experience.