This challenge is prevalent in the real world, especially as the world has shifted attention to the health sector recently. Automating tasks such as sorting patient data could save time and resources that could be used for larger problems. We decided to take on this challenge because it is a practical project and stimulates the brain.
What it does
It takes data as a CSV file of patients, including their address, date of birth, etc and groups multiple entries that are actually the same patient.
How we built it
Build on the ASP.NET framework, we examined patterns in the test CSV, finding common input errors and occurrences. We then used our findings to construct an algorithm that took advantage of these findings to most accurately predict which inputs were for the same patient and differentiated false-positives.
Challenges we ran into
Figuring out an algorithm that would accurately cover specific cases was difficult. It was common to cover specific input errors, but more cases subsequently had to be accounted for.
Accomplishments that we're proud of
Creating an algorithm that is fully functioning and groups patients properly, and accurately deducing which inputs are the same patient, despite of input typos and missing information.
What we learned
How to connect SQL server with ASP.Net, and filling databases with csv data as well as parsing csv data.
What's next for Patient Matching
We hope to continue to improve the accuracy of our algorithm until it can achieve near 100% accuracy and be used at real hospital and businesses that may need to match clients with their respective datas.