BoilerMD

Utilizing healthcare datasets, random forest and logistical regression algorithms, BoilerMD is capable of revolutionizing free, universal healthcare by offering vividly accurate medical assistance to the general public. Through a webpage, the user selects a disease to diagnose, and then enter data to help the algorithm arrive at a diagnosis. Using our custom forest regression models, BoilerMD outputs its diagnoses at high accuracies (up to 97%!).

This tool allows for greater precision in diagnosing the currently offered diseases as the models are inherently more precise than medical professionals. Moreover, the larger the user database and longer the period since deployment, the more accurate the model becomes with time. It simply only gets better. This tool also offers the opportunity for large data analysis; for example, a local government could use hospital data to detect abnormal kidney diseases or obesity in the region, allowing them to put into effect policies that would help the community.

We were inspired to create this project after a recent spike in the so-called “Purdue Flu” which affected thousands of students. Through BoilerMD, we are starting this initiative of large-data AI healthcare services.

In the process of making BoilerMD, we experimented with neural-networks and logistic regressions. Through optimization, we determined that random forest regression algorithms resulted in the greatest accuracy.

We plan to continue improving BoilerMD by refining its design and adding more available diseases to be diagnosed, in our vision to secure reliable, affordable healthcare for all.

Thank you for your time!