Contributors: Arav Parikh, Kaitlyn Bedard
Clinical decision-making (i.e. the formulation of diagnoses) is a critical part of what doctors do on a day-to-day basis; however, it is a rather demanding task given that it requires them to have an in-depth understanding of disease descriptions and identify hidden patterns among the symptoms. While there is no doubt that these doctors possess this knowledge, the time it takes to recall such information detracts from their ability to spend time and directly work with their patients to achieve better clinical outcomes. With the rise of AI/ML models, though, the automation of the task of diagnosis prediction is not only readily achievable but perhaps even preferable in the sense that it might allow patients to receive diagnoses without directly consulting a doctor, thereby reducing undue strain on the healthcare system and saving doctor visits for treatments and emergencies. In this repository, we explore past datasets and models used for this task, identify areas of augmentation, and describe a chosen implementation and examine the results.